Harvey, the A.I. startup set out to automate the legal industry, is courting the sports world to reach new customers. The OpenAI-backed company, which just closed a $550 million funding round this week, is the first official “legal assistant” to the U.S. Open, securing signage at Arthur Ashe Stadium and select other courts, plus virtual on-court placement in international broadcasts. The tournament runs through this Sunday.

“The U.S. Open’s fan base, global reach and remarkable brand make them the perfect partner,” Harvey CEO Winston Weinberg said when announcing the partnership in March.

Legal software may seem an unlikely sell to tennis fans. But Harvey sees a business audience in the crowd, with potential customers watching from the stands and at home.

“CEOs, boards, lawyers and adjacent teams all attend and tune in to the tournament,” Mali Robertson, Harvey’s director of brand partnerships, told Observer in an email. She described the U.S. Open as the company’s first major sports partnership.

Companies that sell to other businesses love the chance to meet prospective customers and other business leaders in person, said T. Bettina Cornwell, a University of Oregon marketing professor who studies corporate sponsorships.

“If there are 10 or 20 other business leaders they might meet during one event, that holds so much more potential than cold calling for interest,” Cornwell told Observer. Other sponsors can become prospective customers, she said, while hospitality gives companies a way to thank existing clients and reward employees.

“Building Harvey’s brand not just with lawyers who are tennis fans but also with a broader audience becomes imperative” as the company grows, Robertson said.

The U.S. Open’s other sponsors include Oura, American Express, J.P. Morgan, IBM, Deloitte, Rolex and Emirates, among others.

Harvey looks beyond Big Law

Weinberg knows the legal market from the inside. In 2022, he was a junior associate at O’Melveny & Myers when he began testing OpenAI’s models on legal questions with his roommate, Gabe Pereyra, then a researcher at Meta. The two co-founded Harvey, which received its first check from the OpenAI Startup Fund.

Investors have kept the money flowing. In March, Harvey raised $200 million at an $11 billion valuation in a round co-led by GIC and Sequoia. This week’s financing, co-led by Diffusion and Lightspeed Venture Partners, brings its total funding to more than $1.5 billion. Harvey is now valued at $15.5 billion.

Harvey says its software is currently used by more than 200,000 lawyers at more than 2,400 organizations, up from more than 100,000 lawyers at 1,300 organizations in March. The customer list includes law firms such as Paul, Weiss and A&O Shearman, PwC UK and private equity giant KKR.

Despite being a B2B company, Harvey’s marketing push has heavily focused on reaching a mass audience. In June 2025, it launched a campaign on New York streets and in the subway, challenging lawyers with the line “you didn’t go to law school to push papers.” In February this year, Harvey partnered with Gabriel Macht, the Suits star whose fictional legal shark, Harvey Specter, inspired the company’s name. The partnership coincided with the launch of Harvey’s Instagram account.

Its sports deals have also tracked its geographic ambitions. A March partnership with the Dallas Mavericks came as Harvey prepared to open a Dallas office in April. Later in March, it struck a deal with Paris Saint-Germain ahead of a planned office opening in the French capital. In addition to stadium branding and customer experiences, the agreement included a commitment by the soccer club to use Harvey in its legal department.

Other A.I. and SaaS companies are also turning to sports to bolster their image. IBM’s relationship with the United States Tennis Association dates back to 1992. Today, its technology powers match analysis and an A.I. chatbot on the U.S. Open’s website and app.

In February, Anthropic partnered with the Atlassian Williams F1 Team, putting Claude branding on cars and team gear and applying the A.I. assistant to race strategy, car development and operations. That month, Anthropic also ran Super Bowl commercials promoting Claude as an ad-free alternative after rival OpenAI began inserting ads for some ChatGPT users.

Lululemon turned everyday yoga wear into a status symbol. Canada Goose did the same for winter coats. Now, two of Canada’s best-known fashion exports face a test of their staying power, particularly in the U.S. Both companies’ stock prices are down more than 80 percent from their peaks. Alo Yoga, Vuori and other athleisure rivals are challenging Lululemon, while Canada Goose is trying to stretch its appeal beyond the parka. Their competition for American shoppers is also unfolding amid an escalating U.S.–Canada trade dispute. U.S. duties imposed in August hit Canada Goose’s Canadian-made apparel, while Lululemon’s largely Asian-made products face other U.S. import tariffs.

Their U.S. sales were already under pressure before the latest tariff escalation. In the most recent quarter ended August, Lululemon’s global sales fell 4 percent, weighed down by an 8 percent decline in the U.S. On the Sept. 3 earnings call, CFO Meghan Frank reported a roughly 20 percent drop in leggings sales and acknowledged “pressure on brand heat and sentiment.”

Canada Goose is weathering a similar slowdown. In its April–June quarter, U.S. revenue dropped 19 percent year-over-year, and global comparable sales fell 3.2 percent, even as overall global sales grew 10.3 percent.

For both brands, the existential question is whether the products that captivated millennials can retain their cachet with a new generation of shoppers.

Lululemon’s familiarity problem

Founded in 1998 in a Vancouver studio that doubled as a yoga space, Lululemon defined and rode the athleisure wave throughout the 2010s. Its stock price peaked at $510 in late 2023 (five times where it trades today) before momentum began to stall in 2024.

To a degree, Lululemon has “gone off the boil” in the U.S., Neil Saunders, managing director of GlobalData Retail, told Observer. While the brand remains massive and profitable, lackluster collections are discouraging repeat purchases and pushing shoppers toward younger rivals.

“Lululemon is not a brand people hate. It’s just a brand that people have become a little bored with,” Saunders said. “Lululemon is not recruiting younger shoppers at the pace it should be.”

The shift is increasingly visible online. Over the past three months, Lululemon’s average weekly U.S. search volume fell 8 percent from a year earlier, according to Trendalytics. By contrast, searches for Vuori rose 11 percent, while Alo Yoga’s average weekly search volume surged 47 percent.

Social media data tells a similar story. Alo’s TikTok following grew 3.1 percent over the past month compared to just 1 percent for Lululemon, while Alo’s Instagram engagement jumped 96.3 percent as Lululemon’s dropped, according to Trendalytics.

“Across the board, Alo is showing the strongest growth and appears to be gaining ground on its competitors,” Cate Khan, CEO of Trendalytics, told Observer.

Adding to the pressure, Lululemon founder Chip Wilson, who left the board in 2015 but remains a major shareholder, has publicly criticized current management. In January, he declared that Lululemon had “lost its cool” after the company briefly paused North American online sales of its $108 leggings following complaints that they became sheer when bending or squatting. Rather than altering the design, Lululemon advised customers to size up and wear seamless, skin-tone underwear—an awkward fix for a brand built on technical performance.

That misstep reinforced perceptions that Lululemon was “out of touch” with Gen Z and younger millennials, said Shawn Grain Carter, an associate professor of fashion business management at the Fashion Institute of Technology. Student debt, inflation, and tariffs—coupled with “disappointing styles”—have created “a serious test of their consumer loyalty,” she told Observer.

Lululemon’s own market research shows that shoppers are clamoring for more “new and differentiated products,” CFO Frank told analysts last week. To regain traction, the company is accelerating product development and doubling down on creator content and community events. Looser silhouettes, such as the Groove Wide-Leg and updated Dance Studio pants, are performing well, but have not yet offset the broader dip in leggings.

“We remain confident these investments will help to reignite our sales trends over time as we continue to elevate our product and marketing execution,” Frank said on the earnings call. Steering that turnaround now falls to Heidi O’Neill, a longtime Nike executive who stepped in as Lululemon’s CEO this week (Sept. 8).

Canada Goose looks beyond the parka

Canada Goose, which launched in a Toronto warehouse in 1957, turned its recognizable sleeve patch into a status symbol for affluent American winters. However, that high brand recognition now comes with built-in limitations.

“The problem with Canada Goose is that it’s an expensive product that has a rather narrow focus,” Saunders said, noting that many shoppers still view it strictly as an outerwear label even as luxury competitors like Moncler and Moose Knuckles encroach on its territory.

In response, Canada Goose has been expanding into rainwear, windbreakers, knitwear and footwear. “We’re successfully evolving Canada Goose into a year‑round luxury brand,” CEO Dani Reiss said in the company’s July 30 earnings release. CFO Neil Bowden acknowledged softer U.S. store traffic, adding that the company plans to increase marketing spend ahead of its peak winter season.

Digital metrics highlight the brand’s complex positioning: Canada Goose’s average weekly U.S. search volume over the past three months fell 14 percent year-over-year, according to Trendalytics, while searches for rival Canadian rival Arc’teryx rose 13 percent.

The deeper challenge lies in defining what consumers want the brand to represent today. Millennials helped make Canada Goose a household name in the U.S., said Thomaï Serdari, clinical associate professor of marketing and director of the Luxury & Retail MBA at NYU’s Stern School of Business. But younger buyers are increasingly wary of overexposed brands and celebrity endorsements, leaning instead toward wellness, experiences and vintage fashion. To this cohort, conspicuous, high-priced logos can read as “cringe,” she noted.

Expanding beyond parkas requires maintaining a clear connection to the brand’s identity. “What makes the new collection distinctively ‘Canada Goose’ other than their logos?” Serdari asked, suggesting its most receptive audience remains existing millennial fans.

Canada Goose maintains that its newer categories reflect the same performance, durability and craftsmanship as its parkas. “The category may change, but the standards remain the same,” the company said in an emailed statement to Observer, noting that apparel, rainwear and windwear accounted for nearly 40 percent of revenue in its latest quarter alongside growth in its core outerwear.

Meanwhile, other Canadian imports are thriving. Vancouver-based Aritzia saw its U.S. revenue jump 54.5 percent year-over-year in its latest quarter, with American shoppers now generating roughly two-thirds of its net sales.

While brand recognition is locked in for both Lululemon and Canada Goose, their hardest task ahead is giving shoppers a fresh reason to buy their next pair of leggings or parka.

When the U.S. Open main draw begins Sunday in Flushing Meadows, players will settle into the usual rhythm of matches, practice courts and media rounds. This year, they’ll also be offered Oura Rings. Under a new five-year agreement with the U.S. Tennis Association, the Finnish smart-ring maker has become the USTA’s first wearable partner. The deal comes as Oura prepares for a closely watched IPO and faces the challenge of proving it is more than a trendy tech accessory worn by celebrities and athletes.

Oura confidentially filed to go public in May. The listing could happen as soon as September and would seek $3 billion in fresh capital at a $16 billion valuation, according to Bloomberg. Oura was last valued at $11 billion, after raising more than $900 million in a Series E round in October 2025.

Oura’s $6-a-month subscription unlocks detailed sleep data, heart-rate trends and other health insights in its app, turning a ring purchase into a source of recurring revenue. A public filing could reveal how much of the company’s growth comes from membership fees rather than one-time ring sales—and, if disclosed, how long customers stick around.

Oura’s Finnish roots and rapid growth

Co-founders Petteri Lahtela, Kari Kivelä and Markku Koskela launched Oura in 2013 in Oulu, Finland, drawing on engineering experience from Nokia and Polar, the Finnish heart-rate-monitor maker.

In 2015, Oura raised $652,000 through a Kickstarter campaign, surpassing its $100,000 goal in 15 hours. That year, the company also raised its first $2.3 million in seed funding, led by Lifeline Ventures.

Oura’s leadership changed as it moved from a niche smart-ring maker to a mass-market brand. Lahtela, the company’s founding CEO, transitioned to president and chief innovation officer in 2018, when Harpreet Singh Rai took over as CEO. Rai, who had joined Oura as an investor in 2016, oversaw the launch of its third-generation ring before stepping down in December 2021. In 2025, Rai sued Oura over his compensation package. The case remains unresolved.

COO Michael Chapp served as interim CEO while the board searched for a successor. After a months-long search, the board named Tom Hale CEO in April 2022. Hale had been president of Momentive and previously held roles at HomeAway and Adobe.

Hale was already an Oura user. During a stressful period in 2021, Hale struggled with sleep and later wrote that the ring’s feedback helped improve his mood and overall health. Under Hale, Oura has pushed beyond its early audience of Silicon Valley biohackers, adding women’s health features and placing its rings in retailers such as Best Buy and Target rather than selling primarily through its own website.

Oura has reported rapid growth. By September 2025, the company said it had sold more than 5.5 million rings, including more than 2.5 million since June 2024. Revenue exceeded $500 million in 2024. In November, Hale told CNBC that Oura was on track for $1 billion in sales in 2025 and that 2026 revenue could run “close to $2 billion.”

Those figures will bring familiar public-market questions: how much it costs Oura to win new customers, whether margins will hold as the company expands in retail and how it plans to compete with Samsung and lower-priced rivals.

Oura’s sports pitch

The USTA agreement gives Oura another platform from which to make its case. The company says its members have logged more than 5.5 million hours of tennis, evidence that the ring already has a foothold in the sport.

Under the deal, Oura will offer on-site fittings and recovery education in player areas. It will also have on-court branding at Arthur Ashe and Louis Armstrong Stadiums, including virtual LED messaging visible to broadcast viewers. Beyond the U.S. Open, the company will sponsor the USTA League National Championships and collaborate on health and wellness studies for coaches.

“Tennis is a sport that’s decided as much by what happens off the court—sleep, readiness and recovery—as by what happens on it,” Doug Sweeny, Oura’s chief marketing officer, told Observer. The U.S. Open, he added, lets Oura bring that idea to “millions of people who may never have thought about recovery as part of their own performance and wellbeing.”

The USTA deal is part of Oura’s wider sports push, which Sweeny described as “a core part of our next phase of growth.” In February, Oura became the official wearable of Team USA, with its rings offered to U.S. Olympians and Paralympians during the Milano Cortina 2026 and Los Angeles 2028 Games. In April, it signed an agreement with U.S. Soccer covering all 27 national teams.

This month, Oura added tennis stars Coco Gauff and Taylor Fritz as ambassadors for a campaign spanning digital, print and out-of-home advertising across New York City. Sweeny said many of Oura’s sports relationships began with athletes, coaches and trainers already using the ring. But the company is aiming beyond competitive sports. “Recovery and readiness aren’t just elite-athlete concepts—they’re universal,” he said.

The strategy comes as wearables gain traction. U.S. retail revenue for fitness-tracking devices rose 88 percent in 2025, while smart rings’ share of category revenue climbed to 75 percent from 46 percent a year earlier, according to Circana.

Oura would also enter a crowded and unpredictable market for high-growth technology companies. SpaceX completed a record-setting offering in June, while OpenAI and Anthropic are reportedly considering trillion-dollar listings.

In 2020, a year before stepping down as Amazon’s CEO, Jeff Bezos pledged $10 billion to tackle some of the world’s biggest climate challenges by 2030 and unveiled the Bezos Earth Fund. The Amazon founder called climate change “the biggest threat to our planet,” promising that the money would support “scientists, activists, NGOs” and offer “a real possibility to help preserve and protect the natural world.” Six years later, the fund has awarded at least $2.4 billion across over 371 grants, according to its website. With around three-quarters of the pledge still to be allocated, the fund’s efforts have fallen behind schedule, though it says it remains committed to disbursing the full $10 billion by 2030.

Andrew Steer, former head of the World Resources Institute, became the Earth Fund’s first president and CEO in 2021. Last year, he was succeeded by Tom Taylor, the longtime Amazon executive who served on Bezos’ senior team (known as the S-Team) and once ran the company’s Alexa division. Bezos’s newlywed wife, Lauren Sánchez Bezos, serves as the Earth Fund’s vice chair.

The Earth Fund says it doesn’t “just write checks and walk away.” But how its projects are chosen remains largely opaque. Nor does it accept unsolicited proposals, saying its grantmaking is directed toward programs already identified on its website.

Therefore, the grants themselves offer the clearest picture of the fund’s priorities. While far from exhaustive, the list below includes some of the Earth Fund’s largest and most noteworthy donations to date.

Phoenix Species Project: $100 million pledged in 2026

In 2021, actor Leonardo DiCaprio and biologist Wes Sechrest met with Bezos and Sánchez to discuss various conservation topics. Those talks ultimately produced the $200 million Phoenix Species Project, unveiled this year. The effort aims to recover critically endangered species across 30 countries. More than 31,000 species assessed by the International Union for Conservation of Nature are categorized as endangered, critically endangered or extinct in the wild. Phoenix will concentrate on 100 of them—a fraction of the global crisis, but a test of whether sustained funding can do more than slow their disappearance.

The Earth Fund is supplying $100 million, while Re:wild, co-founded by Sechrest, has committed the other half with backing from DiCaprio, Dax Dasilva’s Age of Union and the Todd Graves Family Foundation. Graves, the billionaire founder of Raising Cane’s and a friend of DiCaprio’s said the actor’s work with Re:wild inspired him “to help in a meaningful way.”

From that small circle of wealthy donors, the money will flow through more than 100 partners, including Association Fanamby in Madagascar, Bush Heritage Australia and the Amphibian & Reptile Conservancy in North Carolina. The roster spans mammals, birds, amphibians, reptiles, fish, invertebrates and plants, from Australia’s elusive night parrot to Mexico’s Charco Azul pupfish. Others include the Hickory Nut Gorge green salamander, found only in a North Carolina canyon, and South Africa’s “fuzzy belly button,” a succulent poached for the ornamental-plant trade. 

Each species will receive funding over five years, freeing conservationists from the shorter grant cycles that can disrupt recovery work. Organizers say Phoenix will track population recovery and reductions in threats, but the newly launched project is still too early to have produced measurable results.

The Opening Round: $791 million pledged in 2020

At $791 million, the Earth Fund’s first major outlay remains its largest. The oddly specific total covered 16 grants, but nearly two-thirds of the money was concentrated among five organizations: The Environmental Defense Fund, Natural Resources Defense Council, The Nature Conservancy, World Resources Institute and World Wildlife Fund each received $100 million. 

The choices also foreshadowed who would run the fund. World Resources Institute, then led by Andrew Steer, received $100 million over five years for satellite-based land monitoring and an initiative helping U.S. school districts and others plan and finance electric buses. A few months later, Steer left WRI to become the Earth Fund’s first president and CEO.

The grants have since produced a mixed record. WWF said that by late 2025, its mangrove program had protected or restored more than 2.4 million acres across Colombia, Fiji, Madagascar and Mexico. EDF, meanwhile, put part of its grant toward MethaneSAT, a satellite built to detect methane emissions from oil and gas operations. It collected data for roughly a year before its team lost contact with the satellite in 2025. EDF said the broader effort would continue, including through MethaneAIR, a sensor-equipped aircraft.

Greening America’s Cities: $400 million pledged in 2023

Sánchez Bezos is the public face of Greening America’s Cities, which focuses on creating cooler, healthier neighborhoods. In 2023, she unveiled the $400 million program at Los Angeles’ Pacoima Wash alongside then-White House senior adviser John Podesta, Sen. Alex Padilla and other leaders. As extreme heat makes the cooling effects of trees and parks more consequential, access to green space is becoming a form of urban infrastructure.

By mid-2026, the Earth Fund had deployed $150 million through 39 grants in 11 cities, supporting local nonprofits and community-led projects, often in partnership with city governments. In Los Angeles, Pacoima Beautiful, a woman-led environmental justice group founded by local mothers, received $3.5 million to help turn the Pacoima Wash from a flood-control channel into a greenway. In Chicago, Grow Greater Englewood is using its $3.8 million to help convert more than 10 acres of polluted land into an agricultural-eco district and nature trail.

When the Earth Fund announced a second $100 million round of the total commitment in July, Sánchez Bezos recalled taking the bus after school to her grandmother’s Albuquerque home and spending hours in a nearby park. “That’s what we’re trying to give people with this program,” she said. “Not just green space. A place that feels like yours.” 

Sustainable Protein Network: $100 million announced in 2024

The Earth Fund’s alternative-protein push also began with Sánchez Bezos, who unveiled the first $60 million at Aspen Ideas: Climate in 2024. Rather than backing consumer brands, the fund committed nearly all of the eventual $100 million to university research aimed at making plant-based, cultivated and fermented proteins cheaper, better-tasting and easier to produce.

Steer, then the Earth Fund’s president and CEO, framed it as a way to slow the expected rise in meat production without asking everyone to give up meat. “We don’t all need to become vegetarians,” he said at a 2024 Council on Foreign Relations event. The Good Food Institute estimates global meat demand could rise by more than 50 percent by 2050.

The commitment includes $30 million each for Bezos Centers for Sustainable Protein at North Carolina State University, Imperial College London and the National University of Singapore, plus a $9.8 million grant to the Good Food Institute, which helped scope and select the centers. In Singapore, Professor Zhou Weibiao and other researchers are developing what the center calls “ultimate proteins,” including through biomass fermentation that uses tofu-processing waste to feed microalgae. 

The Brazilian Amazon: $100 million awarded between 2023 and 2025

Since 2023, the Earth Fund has awarded $100 million across at least 30 grants in the Brazilian Amazon, betting that the rainforest stands a better chance when its residents can make a living from it. 

Cristián Samper, the fund’s nature lead and a former Wildlife Conservation Society CEO, has described the work as finding “alternative paths to prosperity” for people living in the Amazon. 

On the ground, that includes a $4.2 million grant supporting Cooperacre and SOS Amazonia. The cooperative is adding equipment and infrastructure to expand the sustainable production of rubber, fruits, nuts, oils, coffee and cacao harvested from protected reserves.

Along the Rio Negro, the Earth Fund is also backing FOIRN, the region’s Indigenous federation, to strengthen land governance across 13.5 million hectares of forest, helping communities make collective decisions in a region threatened by logging, mining and other land clearing pressures. 

A broader $30.9 million package was awarded to Nia Tero, Rainforest Foundation Norway, Rainforest Trust and the Wildlife Conservation Society, working with local partners. Their projects range from strengthening Indigenous land governance to securing new protected and conserved areas.

Tennis legend Roger Federer earned nearly $131 million in prize money over his 24-year career. But one of his most lucrative moves came off the court. In 2019, he invested in On, the Swiss running-shoe company he joined as what the brand calls a “co‑entrepreneur.” On has never disclosed the terms, but previous reports said Federer invested roughly $50 million for an estimated 3 percent stake. At recent valuations, that position would be worth about $310 million.

It’s been a tough year for the sneaker company, however. Its share price is down nearly 40 percent in 2026 and plunged earlier this month after an alarming quarterly earnings report. Forbes estimated that Federer’s fortune fell by at least $52 million in a single day, briefly pushing him below the billionaire threshold.

On’s famously lightweight sneakers are still selling well and remain highly profitable. Second‑quarter sales rose 21.6 percent to $1.05 billion, while gross margin reached 65.4 percent. However, investors are worried about cooling demand in the Americas, slower wholesale growth, and a reduced full‑year forecast.

For Federer, a large share of his fortune now rests on whether On can keep pace with Wall Street’s expectations.

Federer was already one of the sports world’s most successful pitchmen. Forbes estimates that endorsements, appearances and other business ventures brought him roughly $1 billion during his playing career (before taxes and agents’ fees). After leaving Nike in 2018 for a reported 10‑year, $300 million apparel deal with Uniqlo, he found himself without a footwear sponsor. His wife, Mirka, wore On’s shoes, and Federer began noticing them everywhere in Switzerland.

On’s founders—former Ironman champion Olivier Bernhard and branding executives David Allemann and Caspar Coppetti—reached out. Federer joined not just as an investor but as a collaborator. According to On’s IPO filing, he spent “many days” in its Zurich lab working on both his namesake sneaker franchise and a high‑performance tennis shoe.

The first commercial result was The Roger Centre Court 0‑Series, a $250 lifestyle sneaker launched in 2020 as a limited run of 1,000 pairs. Later, the Roger Pro, a performance tennis shoe, was introduced, now priced at $220 and engineered for the lateral movements of competitive play.

On has since built a tennis roster spanning well‑known stars and younger prospects, including former world No. 1 Iga Świątek, U.S. Open semifinalist Ben Shelton and rising Brazilian player João Fonseca.

Since On’s 2021 IPO, annual sales have climbed about fourfold, reaching $3.7 billion by 2025. Over 90 percent of its revenue is from shoes. Its sneakers are particularly popular with younger consumers, thanks to patented technology and celebrity endorsements.

Customers under 34 now represent more than one‑third of On’s customer base, executives said on an earnings call this month, with the $170 Cloudtilt and other lifestyle footwear proving especially popular. Earlier this year, a footwear and apparel collection created with Zendaya and her longtime stylist Law Roach sold out faster than expected, with women under 34 accounting for 60 percent of buyers, according to the company.

On shoes are known for their “running on clouds” feel, enabled by the patented CloudTech cushioning—using hollow pods on the sole that compress during landing and lock firm for a springy push‑off. Part of On’s next act rests on LightSpray, a process in which robots create the upper portion of a shoe in about three minutes, largely eliminating the usual cutting, stitching and gluing. Production sites in Zurich and South Korea have increased On’s capacity thirtyfold this year. And demand has followed: co-founder and co-chairman Caspar Coppetti said the $280 LightSpray Cloudmonster 3 Hyper shoes sold out within hours in parts of Asia and generated nearly 20 percent of footwear sales during the opening week of On’s Boston store.

“We are deliberately engineering for the multi‑decade value of a premium brand,” co‑founder and co‑chairman David Allemann told investors on an earnings call this month.

On is expanding, too, recently opening stores in São Paulo, Copenhagen, and Stockholm, with Sydney next. But that growth comes amid leadership turnover. Marc Maurer left his co‑CEO role in 2025, while longtime financial chief Martin Hoffmann stepped down as CEO in May, returning Allemann and Coppetti to daily management. Their challenge is to keep On growing without diluting the premium image that made Federer’s investment so valuable.

Universities are racing to secure the resources needed for serious A.I. research, but competing with deep-pocketed private labs requires more than academic talent. Increasingly, wealthy donors and tech leaders are stepping in to help. Their gifts are funding new colleges and research centers, opening access to A.I. infrastructure and supporting new academic programs. Philanthropists are even helping build the country’s first A.I.-native hospital, as universities play a larger role in shaping how the technology is developed and applied.

Here is a look at nine major gifts driving academia’s next era.

Michael and Susan Dell: University of Texas at Austin, $750 million

Michael Dell arrived at the University of Texas at Austin in the early 1980s, planning to study medicine. Instead, he began selling personal computers from his dorm room and dropped out to build Dell Technologies. Four decades later, he and his wife, Susan, are helping UT Austin build a medical center designed explicitly for the A.I. age.

Their $750 million commitment, unveiled in April, will establish the UT Dell Campus for Advanced Research and its centerpiece, the UT Dell Medical Center, which the university calls the country’s first “A.I.-native” hospital.

With a new facility, “you can design it from the start with data and computing and A.I. built in,” Dell told CNBC. The commitment also supports scholarships, student housing and the university’s supercomputing center, while pushing the couple’s lifetime contributions to UT Austin past $1 billion. Construction could begin as early as September, with the medical center scheduled to open in 2030.

Mark and Mary Stevens: University of Southern California, $200 million

Venture capitalist Mark Stevens earned three degrees from the University of Southern California before joining Sequoia Capital, where he was part of the team behind early investments in Google and Nvidia. Today, he sits on Nvidia’s board and is betting heavily on A.I.’s future.

In May, Stevens and his wife, Mary, gifted $200 million to USC to embed A.I. research across disciplines from medicine to business to the arts.

“With A.I., we’re only in the first inning,” Stevens told The Los Angeles Times, predicting the world a decade from now “will be unrecognizable to us.” The funding will help recruit A.I. researchers working on areas such as new therapeutics, cybersecurity and human creativity. It builds on the couple’s earlier USC gifts of $22 million in 2004 and $50 million in 2015.

USC is also launching a Bachelor of Science in A.I. this fall, adding to more than 30 A.I.- and computing-related programs.

Jensen and Lori Huang: Universities and nonprofit research institutes, $108 million in A.I. compute

Nvidia CEO Jensen Huang and his wife, Lori, met as engineering students at Oregon State University. Their charitable foundation is now helping researchers access the expensive computing infrastructure powering the A.I. boom.

In May, Nvidia disclosed that the Jen-Hsun and Lori Huang Foundation had agreed to purchase GPU computing time from CoreWeave for universities and nonprofit research institutes. According to the filing, $108.3 million in computing resources had been donated so far for open science and A.I. research.

Rather than cash, the gift provides raw compute. CoreWeave operates cloud infrastructure built on Nvidia technology, while the chipmaker invested $2 billion in the company in January.

The Huangs have backed university computing before: in 2022, they gave Oregon State $50 million toward a research complex housing an Nvidia‑powered supercomputer.

University of Wisconsin–Madison’s Catalyst Collective: $100 million

Tech leaders and major donors joined forces to finance UW–Madison’s first new college in more than 40 years.

The Catalyst Collective, including Perplexity co-founder Andy Konwinski, former Cisco CEO John Morgridge, Intuit co-founder Scott Cook, Doximity co-founder Jeff Tangney and software giant Epic Systems, committed $100 million to launch the College of Computing and Artificial Intelligence, which opened July 1.

The funding will help recruit faculty, expand research and build advanced computing infrastructure. UW–Madison plans to add 50 faculty positions through the college, bringing its broader wave of new A.I.-related hires to more than 100. The college combines computer sciences, statistics and the Information School, with priorities including trust, fairness, privacy, environmental sustainability and the future of work.

“A.I. is already reshaping society,” Dean Remzi Arpaci‑Dusseau said. “In moments of major change, universities have a responsibility to engage, not stand on the sidelines.”

Sanford and Joan Weill: Stanford Medicine and the University of California, San Francisco, $100 million

Former Citigroup chairman Sandy Weill and his wife, Joan, are bringing together two of California’s top cancer research centers.

In July 2025, the Weill Family Foundation announced a $100 million matching gift to establish the Weill Cancer Hub West, a collaboration between Stanford Medicine and the University of California, San Francisco. Both institutions pledged to raise an additional $50 million each, bringing total funding to $200 million over a decade. “Good things happen when people collaborate,” Weill said in Stanford’s announcement.

Early projects include building an A.I. model that integrates clinical and biological data to inform personalized cancer treatments. Another project will use A.I. to analyze millions of immune cells and improve therapies for treatment-resistant tumors. Another team is developing a CRISPR-based method to engineer cancer-fighting immune cells inside the body.

Reed Hastings: Bowdoin College, $50 million

Netflix co-founder Reed Hastings traces his interest in A.I. back to a conversation with a Bowdoin professor in 1982. In 2025, he returned to his alma mater with a $50 million gift, the largest in Bowdoin’s history, to establish the Hastings Initiative for AI and Humanity.

Rather than focusing narrowly on technical development, the initiative examines how A.I. could reshape ethics, work, creativity, relationships and inequality. “Our moral‑ethical system improvements need some bolstering,” Hastings told The New York Times.

Bowdoin is using the gift to fund 10 faculty positions across areas, including computer science, philosophy, anthropology, neuroscience and government, and has launched programming on issues ranging from algorithmic bias to A.I.’s environmental costs.

Joe and Rika Mansueto: University of Chicago, $50 million

Morningstar founder Joe Mansueto and his wife, Rika Yoshida, both University of Chicago alumni, gave the university $50 million in April to launch the Mansueto Faculty of Mind and Machine Challenge.

The challenge aims to attract additional donors and generate nearly $200 million in total to recruit and support 20 scholars across fields ranging from medicine and law to science and the humanities. Funding will also support the broader research and educational ecosystem around A.I. The effort builds on an existing UChicago initiative backing faculty projects in areas including oncology, archaeology, public policy and the visual arts.

“With A.I. advancing at an unprecedented pace, the most important breakthroughs won’t come from technology alone,” Mansueto said.

Arnie and Lauren Bellini: University of South Florida, $40 million

ConnectWise co-founder Arnie Bellini and his wife, Lauren, are positioning Tampa as a technology hub dubbed “Cyber Bay.”

In March 2025, the couple gave the University of South Florida $40 million, the largest gift in its history, to establish the Bellini College of Artificial Intelligence, Cybersecurity and Computing, which USF calls the nation’s first named college focused on the intersection of A.I. and cybersecurity.

“Cyberattacks are the new missiles, and A.I. is the arms race of our time,” Bellini said at the time. When the college officially launched in August 2025, 14 newly hired professors joined its faculty, bringing expertise in areas including machine learning, trustworthy A.I. and biomedical applications.

Tom Secunda-led donor group: Binghamton University, $30 million plus $25 million from New York State

Bloomberg LP co-founder Tom Secunda is helping Binghamton University build what it calls the first independent A.I. research center at a U.S. public university.

In January, Secunda and a group of donors committed $30 million to establish the Center for AI Responsibility and Research. The private funding will support the center’s launch, while a $25 million capital investment from SUNY will support a permanent facility and equipment.

“This new center will help create the tools, standards, and talent pipeline the country needs so A.I. is not only powerful, but also secure, transparent, and worthy of public confidence,” Secunda said in a statement.

Binghamton is also an early member of New York’s Empire A.I. consortium, giving universities access to elite computing resources. Faculty and students already lead 10 projects approved to use Empire Alpha, the consortium’s first supercomputer.

As Tesla’s CEO, Elon Musk is practically synonymous with electric vehicles, although his personal history began with gas-powered cars. In 1994, at age 23, Musk paid $1,400 for a used BMW 320i, which he repaired himself and drove for two years. Over the decades that followed, he also owned a 1967 Jaguar E-Type, a Porsche 911 Turbo and a McLaren F1 that cost close to $1 million. Then there was “Wet Nellie.” In 2013, Musk paid nearly $1 million for the Lotus Esprit submarine prop seen in the 1977 James Bond film The Spy Who Loved Me. The car later inspired Tesla’s Cybertruck.

While Musk has never published a complete inventory, interviews and social media posts provide a partial record of the classic and collectible vehicles the Tesla CEO has bought, kept and sold over the years.

Ford Model T (likely 1920s)

Of the two gas cars Musk most recently said he owned, the Ford Model T is perhaps the most unexpected. Musk first disclosed the vehicle in October 2017, writing on Twitter (now X) that he received it from a friend. Several automotive roundups describe it as a 1920 model, but none provides documentation. As of 2024, the antique Ford was still in his collection.

Between 1908 and 1927, Ford built more than 15 million Model Ts. By the early 1920s, more than half of the registered automobiles worldwide were Fords. The automaker’s moving assembly line helped lower the Model T’s price from $850 to $260 — or roughly $5,000 today — putting car ownership within reach of the broader public.

A century later, Tesla adopted a similar strategy. In his 2006 “master plan,” Musk said the company would begin with an expensive sports car and use the proceeds to develop more affordable models. The plan eventually took Tesla from its first production vehicle, the Roadster, to mass-market models like the Model 3 and Y.

In 2025, Musk credited Ford for establishing the foundation for large-scale manufacturing. Responding to an X post in which Ford CEO Jim Farley compared the automaker’s new EV production with its Model T-era assembly line, Musk wrote, “Ford basically invented mass manufacturing of large, complex products. Everyone else copied. Most people don’t know this.”

Jaguar E-Type Series 1 Roadster (1967)

Unveiled at the Geneva Motor Show in 1961, the British sports car paired a sleek, aerodynamic body with a top speed of 150 mph, yet cost only a fraction of its rivals.

Musk’s love for the vehicle began at age 17, when he obtained a book about classic convertibles. “I looked through all, and the one I liked the best was the E-type Jaguar,” he recalled in a 2012 Forbes interview. “I said, ‘Well, if I can ever afford it, that is the car that I want to get.’”

That opportunity came during the early days of Zip2, Musk’s first major startup. He spent a small bonus from an investor on the Jaguar. But the excitement was short-lived. “It was like a bad girlfriend — it kept breaking down on me, causing me all sorts of trouble,” he said. “In fact, it broke down on the way back from the dealer,” Musk told Forbes. 

Yet, he held onto it. In a 2017 post on Twitter, Musk called the vehicle “my first love.” By 2024, it remained one of the only two gasoline cars he owned, alongside the Model T.

Lotus Esprit S1 “Wet Nellie” (1977)

Unlike the other vehicles in Musk’s collection, this one cannot be driven on the road. In 2013, Musk bought the modified Lotus Esprit at an auction through an anonymous bid for £616,000, or about $920,000 at the time.

Decades earlier, the vehicle had been converted into a working submarine for the 1977 James Bond film The Spy Who Loved Me, in which Roger Moore’s 007 drives the car off a pier before it appears to transform underwater.

At the time, the transformation captivated audiences, including a young Musk. “It was amazing as a little kid in South Africa to watch James Bond in The Spy Who Loved Me drive his Lotus Esprit off a pier, press a button and have it transform into a submarine underwater,” Musk wrote in a 2013 statement reported by The Guardian. “I was disappointed to learn that it can’t actually transform.”

Built as a “wet” submarine—meaning its cabin filled with water—the Lotus was piloted during filming in the Bahamas by a retired U.S. Navy SEAL in full scuba gear. Battery-powered electric propellers moved it through the water, while Alka-Seltzer tablets created its trail of bubbles, according to RM Sotheby’s.

After the film, “Wet Nellie” wound up in a storage unit in Holbrook, N.Y. When the lease expired, a Long Island couple bought its contents in a 1989 blind auction for a reported $100 and discovered the damaged but largely intact submarine beneath blankets. It was later authenticated and restored before Musk acquired it at auction in 2013.

Musk later wrote on X that he planned to add a Tesla powertrain and make the vehicle transform “for real.” Years later, when Tesla unveiled the Cybertruck in 2019, he said its design was partly influenced by The Spy Who Loved Me.

BMW 320i (1978)

Before the Bond submarine and even Tesla, Musk’s automotive history began more modestly with a used 1978 BMW. Musk bought the car in 1994 for $1,400, when he was 23. 

The 320i belonged to the E21 generation, the first version of BMW’s 3 Series. Introduced in Europe in 1975 and brought to North America for the 1977 model year, the compact sedan had rear-wheel drive and BMW’s familiar kidney-shaped grille.

Musk said he kept his BMW for about two years, including during the early days of Zip2. Its time with him ended—or at least took a turn—after Musk lent the car to a company intern, who called him to report that one of the wheels “fell off the car,” he told Forbes.

Musk has not said publicly if the BMW was repaired, sold or scrapped, but his later posts on Twitter/X suggest that the 320i has long since left his collection.

McLaren F1, Chassis 067 (1997) 

If the Jaguar E-Type was Musk’s first love, the McLaren F1 was his first extravagance. After Compaq Computer acquired Zip2 for $307 million in 1999, Musk received roughly $22 million and spent around $1 million on the magnesium-silver 1997 F1, one of seven F1 cars imported to the U.S. through the Ameritech program. In a 2018 Twitter post, Musk recalled choosing between a supercar and a house in Palo Alto. “Was no contest,” he wrote. “I bought F1 & a small condo that was much cheaper than the car.”

Designed under Gordon Murray, the F1 was the first production road car built around a carbon-fiber monocoque. Its 627-horsepower BMW V-12 sat behind a three-seat cabin, with the driver positioned in the center and two passengers slightly behind. McLaren produced 106 examples across all versions, including 64 road cars. Musk said he used his regularly. 

In 2000, Musk was driving the F1 on Silicon Valley’s Sand Hill Road with PayPal co-founder Peter Thiel.  The duo was headed to a meeting with Sequoia Capital’s Michael Moritz when Thiel asked, “So what can this do?”

“Watch this,” Musk recalled replying in a 2012 interview with PandoDaily. “I floored it and did a lane change on Sand Hill. The McLaren has no traction control or anything. It only weighs a ton, so it has massive power.”

The car spun into a roughly 45-degree embankment, launching it into the air “like a discus,” Musk said. Neither man was injured, though the uninsured F1 sustained extensive damage. Musk and Thiel hitchhiked to their meeting; Musk later had the car restored and said he eventually sold it at a profit.

The sale was among several efforts to raise cash as Tesla and SpaceX faced financial pressure starting in 2008. “The McLaren is a great car,” he later told Wheels Magazine in 2013. “It’s a work of art, a really beautiful piece of engineering, but I didn’t want people always writing that I have a high-performance gasoline sports car, so I decided to sell it.”

Since then, McLaren F1 values have risen considerably. In December 2025, a different road-going F1 sold through RM Sotheby’s for a record $25.3 million. Another will come to market during this year’s Monterey Car Week. This weekend, the auction house will offer a separate 1996 F1 GTR with a stated value above $35 million.

Known as the “Pop Art F1,” that race-derived prototype spent more than 25 years with Pink Floyd drummer Nick Mason. While not directly comparable with Musk’s F1, the figures show how sharply the market has moved since Musk’s own 1999 purchase.

Porsche 911 Turbo (model year unknown)

After eBay acquired PayPal for $1.5 billion in 2002, Musk began exploring the potential of electric sports cars. Around that time, engineer JB Straubel, who later served as Tesla’s longtime CTO and is now recognized as a co-founder, told Musk about the tzero, a battery-powered vehicle developed by California-based AC Propulsion. Musk then arranged a test drive, according to a 2008 Fortune account.

Musk tried to buy the tzero, but AC Propulsion declined to sell it. The company also rejected his $250,000 offer to convert his Porsche 911 Turbo to electric. Porsche eventually introduced an EV of its own, the Taycan, in 2019; the model remains in production.

The conversation nevertheless led Musk toward Tesla. Tom Gage, then CEO of AC Propulsion, introduced him to Martin Eberhard, who had established Tesla with engineer Marc Tarpenning in 2003. After a two-hour meeting in February 2004, Musk agreed to invest $6.3 million and become Tesla’s chairman. 

A 2012 ForbesLife profile indicated that Musk still owned a Porsche 911 at the time. By October 2017, however, Musk said the E-Type and Ford Model T were the only gasoline cars he still owned.

The Musk Foundation, Elon Musk’s private charitable foundation, had its biggest year yet in 2024, giving away roughly $474 million. Online, however, there was little sign of it. Its bare-bones website lists only five priorities—renewable energy, space exploration, pediatric research, science education and safe A.I. —and offers no staff biographies, application instructions or contact information.

Foundations tied to LinkedIn co-founder Reid Hoffman, Google co-founder Sergey Brin and Nvidia CEO Jensen Huang similarly manage billions of dollars while saying remarkably little about how they operate. Their tax returns reveal where their checks landed, but they rarely explain how recipients got in the door or what the donors hope to achieve. That quiet approach stands in contrast to philanthropies like the Gates Foundation and Bezos Earth Fund, which keep prominent public profiles and publish extensive information about their strategies and grantmaking.

That silence can look secretive. In philanthropy, though, it is not unusual.

“Many philanthropists choose a lower profile to protect their privacy and personal security, manage the volume of unsolicited funding requests or keep the focus on the nonprofit and its mission rather than on the donor,” Gillian Howell, national philanthropy executive at Foundation Source, told Observer.

While visibility can help donors draw attention to causes and inspire others to give, “it’s common for donors to have concerns, and in some cases, less visibility can support greater impact and effectiveness,” Howell said.

Rather than operating like traditional mega-charities with public review boards, these billionaire founders often skip public-facing portals and open applications entirely. Instead, they write multimillion-dollar checks through advisers, personal networks and private tips, often avoiding the flood of pitches that can come with a high-profile presence. 

Musk builds this discretion into his giving philosophy. In 2019, a spokesperson for Musk Foundation told The Guardian that Musk believed true philanthropy was done, for the most part, anonymously. Six years later, Musk returned to the topic on Nikhil Kamath’s “People by WTF” podcast, saying his foundation’s greatest challenge was giving money in a way that was “truly beneficial to people.”

“It’s very easy to give money away to get the appearance of goodness,” Musk said. Achieving the “reality of goodness,” he added, is far more difficult.

Yet, like any tax-exempt private foundation, his organization is legally required to file annual IRS disclosures, offering the public a rare paper trail into where those hundreds of millions actually go.

According to its latest return, the Musk Foundation’s record $474 million payout in 2024 was highly concentrated. About $370 million went to The Foundation, a separate Musk-backed nonprofit behind the new Ad Astra school in Bastrop, Texas, alongside broader plans for primary and secondary education and, ultimately, a university. Its latest return reported less than $1 million in annual expenses and nearly $780 million in net assets, indicating that most of the funding remained with the nonprofit rather than being spent on the current school.

Another $35 million went to Fidelity Charitable, a sponsor of donor-advised funds that makes eventual recipients more difficult to trace. The remainder was divided among dozens of organizations working in healthcare, science, education, affordable housing and disaster relief, from Austin Habitat for Humanity to Code.org and Team Rubicon. Musk was the sole contributor that year, donating Tesla stock then valued at $73.9 million. The foundation’s much larger payout, therefore, drew on assets accumulated in earlier years. While the filing shows his financial interests, it leaves out how recipients were chosen or what Musk expects from them.

Why billionaires prefer to give behind closed doors

Fear of public scrutiny often discourages donors from being transparent — or even moving more capital, Nicholas Tedesco, president and CEO of the National Center for Family Philanthropy, told Observer. A grant that attracts negative coverage “may loom large in the mind of a donor — much more so than the many successful grants that get less public attention,” he said.

There are practical bottlenecks, too. A public presence often invites inquiries from nonprofits, said Katherina “Kat” M. Rosqueta, founding executive director of the University of Pennsylvania’s Center for High Impact Philanthropy. A lean foundation may simply lack the staff to vet and respond thoughtfully to every email.

Staying quiet can spare the nonprofits themselves. “The market for nonprofit funding is incredibly inefficient,” Rosqueta said. “Nonprofits spend a lot of time and energy chasing dollars. No foundation can fund every organization seeking money.” Staying closed can prevent charities from investing scarce hours into applications that will never be read.

Hoffman has been unusually candid about how his closed-door system works. In 2024, his Aphorism Foundation held about $1.1 billion in assets and made over 25 grants totaling roughly $43.4 million. Its tax return lists just two unpaid officers, including himself as president, treasurer and director, and Frank F. Huang, the foundation’s secretary. Aphorism has no public website and accepts no unsolicited proposals.

“It’s never by application. That’s not the way we operate,” Hoffman told Inside Philanthropy in 2024. Instead, he relies on referrals from his network to find people pursuing ambitious, high-risk ideas, treating charity much like a venture capitalist finding founders through introductions.

Aphorism’s two largest awards in 2024—about $10 million each—went to Earth Species Project, which uses A.I. to decode animal communication, and the Probabilistic Computing Foundation, which backs open-source computing research.

“Generally speaking, we don’t seek publicity in any of this stuff,” Hoffman said in the same interview. Asked why Aphorism lacked even a LinkedIn page, he noted that one wouldn’t improve the network referrals reaching his foundation.

Where Hoffman’s giving resembles venture capital, Sergey Brin’s philanthropy is more personal.
Brin’s mother was diagnosed with Parkinson’s disease, and Brin carries a mutation in the LRRK2 gene associated with an elevated risk of developing it.

In 2024, the Sergey Brin Family Foundation held about $4.3 billion in assets and reported $722 million in charitable disbursements. More than $250 million went to the Michael J. Fox Foundation for Parkinson’s Research, while other major grants supported climate, public health and basic science.

Despite that scale, the foundation has no public website and says it supports only preselected organizations. Inside Philanthropy has reported that Brin’s family office, Bayshore Global Management, appears to handle his charitable giving. The foundation’s tax filings also show that outside firms provide administrative and consulting support.

Different vehicles, different visibility

Not every family relies on a single charitable structure. “They may use a donor-advised fund for individual giving, which offers less visibility, but then use a foundation for collective giving, which is inherently more transparent given 990s are publicly available,” Tedesco said. 

Jensen and Lori Huang, for example, saw their foundation’s assets rise from $3.4 billion in 2023 to $9.2 billion in 2024 as Nvidia’s stock surged. (Nvidia’s stock price has risen another 54 percent since then.) It distributed $126 million that year while listing the Huangs as unpaid officers.

Yet roughly two-thirds of that payout was routed through quieter channels: $82.5 million went to the Schwab Charitable Fund (now DAFgiving360), a sponsor of donor-advised funds. These allow donors to claim immediate tax deductions while keeping individual grant recommendations off public records.

The Huangs are not averse to public giving. The couple, who met as undergraduate engineering lab partners at Oregon State University, announced a $50 million gift to the school in 2022 to fund an A.I. and robotics research center housing an Nvidia-powered supercomputer.

Traditional transparency can serve its own purpose. When used thoughtfully, visibility helps donors “promote the causes they care about, establish credibility, attract like-minded partners, build a legacy and lead by example,” Foundation Source’s Howell noted.

The Gates Foundation, for instance, reported $8.01 billion in total expenses in 2024, including nearly $6.9 billion in grants and direct charitable contracts. It maintains a fully searchable public database, publishes detailed program strategies and posts audited financial statements online. Similarly, the Bezos Earth Fund—established through Jeff Bezos’s $10 billion climate pledge—says it has awarded $2.4 billion to date and publicly details its programs, priorities and grant recipients.

More than a century after chemist Eugène Schueller turned a hair-dye formula into the business that became L’Oréal, the beauty giant is still finding new ways to grow, with A.I. as its latest bet.

L’Oréal’s Revenue rose 6.3 percent on a like-for-like basis, a key retail metric, to €11.6 billion ($13.2 billion) in the April-June quarter. For the first half of 2026, revenue increased 5.8 percent to €23.78 billion ($27.1 billion), while net profit climbed 4.7 percent to €3.96 billion ($4.5 billion) and the operating margin reached 21.3 percent. Growth extended across every region and division, from Maybelline’s mass-market makeup to La Roche-Posay’s dermatological skincare and Kérastase’s salon-grade haircare.

CEO Nicolas Hieronimus credited that performance to e-commerce and L’Oréal’s innovation strategy, which now includes a partnership with OpenAI announced in June. “Our innovation engine is firing on all cylinders—and A.I. will help it maintain its pace,” he said in a statement accompanying the results.

Few investors have more riding on that momentum than Schueller’s descendants. His granddaughter, Françoise Bettencourt Meyers, and her family own 34.79 percent of L’Oréal, making them the company’s largest shareholder. L’Oréal shares rose 3.6 percent to €397.15 in Thursday trading, adding roughly €2.5 billion ($2.9 billion) to the market value of the family’s stake. Forbes currently estimates the family’s collective fortune at $95.6 billion.

L’Oréal is turning ChatGPT into a beauty counter

L’Oréal’s e-commerce sales jumped 18 percent to €7.4 billion ($8.5 billion) during the first half of the year, almost twice the growth rate of the online beauty market, according to the company. L’Oréal now sees chatbots as a potential next storefront.

On the earnings call, Hieronimus said younger consumers are “shifting massively” toward large language models for product questions and described beauty as one of the most conversational categories. L’Oréal’s brand recognition, scientific data and professional endorsements could help its products surface in A.I.-generated recommendations, he argued, although the channel remains a “moving target.”

The shift could be lucrative well beyond L’Oréal. McKinsey estimated last year that generative A.I. could create $9 billion to $10 billion in annual value across the beauty industry, with marketing and sales among the largest opportunities.

At VivaTech in Paris in June, L’Oréal outlined a two-pronged collaboration with OpenAI: turning ChatGPT into an interactive beauty assistant while putting its models to work behind the scenes.

On the consumer side, Maybelline plans to bring its virtual makeup try-on into ChatGPT this summer using ModiFace, the augmented-reality company L’Oréal acquired in 2018. L’Oréal is also working with OpenAI to improve how brands, including Lancôme and Kérastase, surface in U.S. ChatGPT conversations. Meanwhile, SkinCeuticals, CeraVe and Garnier are participating in the platform’s advertising pilot, which began in the U.S. in February and has since expanded to multiple international markets.

The partnership does not stop at the digital beauty counter. L’Oréal says its researchers are using GPT-Rosalind, OpenAI’s life-sciences model, to map the skin microbiome and identify beneficial bacteria that could inform future skincare products, starting with La Roche-Posay. OpenAI’s technology will also be integrated into CreAItech, L’Oréal’s internal platform for creating on-brand images, videos and other assets. Hieronimus said the system now produces more than 500,000 pieces of content each month for social platforms.

The A.I. push also extends across L’Oréal’s workforce. The company says 73,000 employees have completed generative A.I. training. In an October internal survey of 4,700 L’OréalGPT users, half said the assistant saved them more than an hour each week.

L’Oréal is not alone in courting shoppers through A.I. assistants. LVMH-owned Sephora launched a U.S. ChatGPT app in March, offering tailored recommendations. Ulta Beauty followed in April with a Google partnership that makes its assortment shoppable through Gemini and A.I. Mode in Search, including checkout for some purchases.

Beauty’s feel-good effect keeps sales growing

Technology may shape how L’Oréal grows next, but its current momentum still rests on a familiar consumer impulse: the desire for indulgence, especially during stressful times.

“The grimmer the economic and geopolitical headline, the more consumers crave an affordable feel-good treat,” Hieronimus said, describing what he calls the “dopamine effect of beauty.”

Haircare and fragrance—two categories that span both mass and prestige—were the fastest-growing segments, up 15.6 percent and 10.3 percent in the first half, respectively. Fragrance also helped L’Oréal Luxe—home to Lancôme, Yves Saint Laurent Beauté and Giorgio Armani Beauty—grow 5.1 percent, nearly twice as quickly as the wider prestige-beauty market, the company said. In contrast, LVMH’s Perfumes & Cosmetics business was flat organically over the same period.

Meanwhile, L’Oréal is widening its luxury offerings through a €4 billion alliance with French luxury giant Kering. The deal, completed in March, added fragrance house Creed into L’Oréal’s portfolio and secured long-term beauty licenses for Kering-owned Bottega Veneta and Balenciaga. Pending regulatory approval, L’Oréal will also take over Kering-owned Gucci Beauty in July 2027, one year earlier than planned. The company hopes to turn Gucci into a multibillion-euro beauty brand, with its first new products expected as early as 2028. 

The Bettencourt Meyers family’s generational handoff

That breadth—from Maybelline and CeraVe to Prada and Gucci—is what has made L’Oréal such an enduring asset for the Bettencourt Meyers family.

Schueller founded the company in 1909 around a hair-dye formula sold to Parisian salons. After his death in 1957, his stake passed to his only child, Liliane Bettencourt, and then to her only daughter, Françoise Bettencourt Meyers, after Liliane Bettencourt’s death in 2017.

The famously low-profile family exercises its influence more subtly than some of Europe’s other luxury dynasties, channeling its power through ownership, board seats and Téthys, its holding company.

That structure underwent a handoff last spring, when Bettencourt Meyers left L’Oréal’s board after 28 years. Her elder son, Jean-Victor Meyers, succeeded her as vice-chairman, while her younger son, Nicolas Meyers, remained a director. Both also sit on the supervisory boards of Téthys and its investment arm, Téthys Invest. Bettencourt Meyers continues to chair Téthys, while her husband, Jean-Pierre Meyers, serves as CEO, according to its website.

The arrangement preserves the family’s influence while leaving day-to-day operations to professional managers. It also leaves Schueller’s great-grandsons firmly in place as L’Oréal pursues its next generation of consumers.

LVMH is showing its clearest signs yet of emerging from luxury’s two-year downturn, lifted by stronger U.S. spending and fresh wealth creation, even as the Arnault family faces renewed scrutiny over who will succeed 77-year-old chairman and CEO Bernard Arnault.

The world’s largest luxury-goods company said Monday (July 27) that revenue rose 3 percent organically to €19.3 billion ($22.2 billion) in the April-June quarter, its best quarterly performance since early 2024. Net profit for the first half of 2026 came at €5.7 billion ($6.5 billion), virtually unchanged from a year earlier, on €38.6 billion ($43.9 billion) in revenue.

The regional picture was mixed, however: U.S. sales rose 6 percent in the second quarter, European spending stalled, and Chinese demand was little changed.

The timing also highlights a tighter grip on the family business. The Arnault family recently crossed the 50 percent ownership threshold in LVMH and now controls roughly two-thirds of voting rights, but Arnault has not named a successor despite having all five children in senior roles.

What’s driving LVMH’s business?

LVMH’s portfolio spans more than 75 brands, including Louis Vuitton, Dior, Tiffany & Co. and Sephora. Its financial results are a closely watched gauge of the global luxury goods market, which Bain & Company expects to reach roughly $420 billion this year.

In a statement accompanying the results, Arnault attributed the second-quarter uptick to the strong debut of new Dior collections by creative director Jonathan Anderson, who joined the brand last year, the opening of new Louis Vuitton stores in Beijing and Seoul, and continued demand for signature jewelry collections at Tiffany and Bvlgari. He also cited growth at Sephora and a recovery in champagne and cognac.

The core Fashion & Leather Goods division, which generates nearly half of LVMH’s revenue, edged up 1 percent, reversing seven consecutive quarters of declines. Watches & Jewelry grew 11 percent, driven primarily by high-end jewelry sales at Tiffany and Bvlgari.

Regional performance reflected shifting economic trends. U.S. sales grew 6 percent in the quarter, up from 3 percent in the previous quarter. LVMH linked its strong U.S. performance in part to new wealth creation, with Reuters pointing to the A.I. boom. “Where wealth is created, consumer appetite for luxury and for our products in particular is strong,” LVMH CFO Cécile Cabanis told analysts during Monday’s earnings call.

Luxury brands still need middle-class customers

Analysts have questioned whether years of price hikes have placed many luxury products out of reach for middle-class shoppers. Cabanis said LVMH must serve “both the very exclusive client and the aspirational, because we need to continue to recruit.”

To sustain that appetite across client tiers, LVMH is balancing high-end offerings with more accessible entry points. Louis Vuitton’s Speedy P9 line, aimed at top-tier clients, starts at $9,450 and can exceed $15,000, with special editions priced even higher. By comparison, newer releases such as Squire bags ($2,470 to $2,860) and Multipass models ($2,230 and $3,350) target “aspirational” consumers, Cabanis said.

LVMH shares slipped 1.5 percent on Tuesday and remain down roughly 30 percent for the year, as analysts questioned whether the earnings numbers are sufficient to signal a broader turnaround. Bain & Company expects the global personal luxury goods market to grow just 2 to 4 percent this year.

The Arnault family tightens its grip

Arnault did not inherit a luxury empire. In 1984, he began building one when he took over the struggling Boussac group, then the owner of Christian Dior. After Louis Vuitton and Moët Hennessy merged three years later, he accumulated a controlling stake in the newly formed LVMH and became its chairman and CEO in 1989. Over the next three decades, he systematically assembled a portfolio, absorbing brands like Sephora, Fendi, Bvlgari and Loro Piana, culminating in the $15.8 billion acquisition of Tiffany & Co. in 2021.

The family’s equity control has consolidated further in recent years. As of June 30, the Arnaults held 50.21 percent of LVMH shares and controlled 66.39 percent of its voting rights.

To safeguard the group against external threats, Arnault restructured his family holding company, Agache, into a joint-stock partnership in 2022. Under the agreement, each of his five children—Delphine, Antoine, Alexandre, Frédéric and Jean—holds a 20 percent stake, according to regulatory filings. All five hold executive positions across key Maisons and operating divisions, with four serving on boards and two on the Executive Committee.

Their father, meanwhile, has shown little urgency to name a successor. Last year, shareholders approved a bylaw change effectively allowing Arnault to remain chairman and CEO until age 85. When a shareholder asked about succession at LVMH’s annual meeting in April, he suggested revisiting the question “in seven or eight years.”

That issue flared up publicly just ahead of earnings. On Sunday, Arnault issued a rare statement pushing back against a multi-part series in the French newspaper Le Monde examining his business empire, whose final installment alleged growing friction among his children. He dismissed the claims of a family feud, writing that luxury “is one of the rare sectors that thinks in generations rather than quarters.” His children, he wrote, “run Houses, build teams, make decisions.” Anyone “betting on a family rift to sell newspapers,” he added, would be “waiting for a long time.”

Balancing algorithms and artisanship

While the Arnault family maintains firm structural control, LVMH’s next phase hinges on balancing heritage with innovation: integrating technology into a business built on scarcity and craftsmanship. A.I. promises speed and scale, but those qualities can sit uneasily beside the slow, hands-on work that gives luxury goods much of their value.

That duality was evident across the quarter. In the earnings statement, Arnault pointed to Les Journées Particulières, an October event that will open the group’s normally private workshops and historic ateliers to the public, showcasing what he described as “the invaluable expertise of our craftspeople and the rich heritage of our Maisons.”

At the same time, digital tools are playing a growing role in creative and operational work. At VivaTech in June, Arnault said he takes part in weekly design sessions using generative A.I. to develop product concepts.

Across LVMH, A.I. is being used to improve client engagement, demand forecasting and product development, Cabanis said. She suggested the technology could also shorten time to market and lower inventory, creating efficiencies that can be reinvested. LVMH would not cut costs at the expense of client service, she added. Louis Vuitton designers are already using A.I. to test colors and visualize materials, while Sephora recently launched an app inside ChatGPT that offers U.S. shoppers product recommendations.

Nvidia built its empire on chips powering generative A.I., and that engine is still surging. The company posted a record $81.6 billion in quarterly revenue, up 85 percent year over year. But with a $5 trillion valuation that prices in years of growth, CEO Jensen Huang is under pressure to prove expansion can extend beyond the U.S. data-center boom. His latest pitch centers on Japan, where the government and major industrial players are gearing up to deploy A.I. in robots and factories.

During a two-day visit to Tokyo last week, Huang rallied Japan’s manufacturing giants, robotics innovators and software developers behind “physical A.I.,” the technology that allows robots and other machines to perceive their surroundings, make decisions and act in the real world. Japan, Huang argued, holds a natural advantage for such a shift. “Japan has historically been very good at precision manufacturing and very large-scale manufacturing, but now we have A.I. You can combine the two technologies and create robotics,” he told reporters during a July 15 event.

Japan has strong incentives to embrace the partnership. It boasts world-class industrial hardware but faces a worsening labor shortage. Prime Minister Sanae Takaichi, who has made semiconductors and A.I. central to her growth agenda, is pushing to pair that hardware strength with more advanced software. Japan accounts for about 70 percent of the global industrial robot market but just over 10 percent of the service robot market, according to government data. Its updated strategy targets more than 30 percent of the emerging A.I.-robotics market by 2040, representing $133 billion in business.

During his Tokyo visit, he pitched his vision to the country’s industrial elite. At a pub in Tokyo’s Kanda district, Huang gathered more than 30 executives from 16 major firms—including Tokyo Electron, Panasonic and Mitsubishi Electric, among others—to discuss how Japan’s semiconductor supply chain could support an A.I.-led expansion. Over lunch, he also met with executives from Fujitsu, Kawasaki Heavy Industries, Fanuc and Yaskawa. According to Nvidia, all four companies are now developing physical-A.I. systems on its platform.

Tokyo’s reception contrasts with Nvidia’s experience in China. Huang visited the country in January and returned in May as part of President Trump’s delegation, trying to rebuild Nvidia’s position in a market squeezed by U.S. export controls and Beijing’s support for domestic chipmakers.

Japan, by comparison, is putting government support behind infrastructure built around Nvidia’s technology. On the second day of his trip, Huang appeared alongside Japanese Industry Minister Ryosei Akazawa to launch the nation’s new government-backed physical A.I. initiative. As part of that kickoff, Huang announced that Nvidia is partnering with Noetra Corp, a Japanese A.I. consortium backed by Sony, SoftBank, Honda and roughly 40 other companies, to build what the chipmaker calls “the world’s first national A.I. infrastructure for physical A.I.”

“Japan must own, improve, secure and deploy Japan A.I.,” Huang said during keynote remarks. “After 15 years of work, physical A.I. is here, the foundation of the next industrial revolution, and it should be made in Japan.”

The project centers on an “A.I. factory” powered by 13,750 Nvidia Vera CPUs and 27,500 next-generation Rubin GPUs. Expected to deliver 140 megawatts of data center capacity, the facility will supply the compute required to train complex physical A.I. models. Construction is scheduled to start in April 2027, with operations coming online in June 2028.

The infrastructure will serve as the computing backbone for FRONTia, a government project supporting multimodal models for physical A.I. Noetra will lead the development of the domestic models running on the network. According to Noetra’s roadmap, the group will prioritize Japanese-language understanding and reasoning before expanding into text, image, video and audio capabilities in 2028. By 2030, it aims to deploy “real-world native A.I.” for robotics and autonomous machinery.

This week, Nvidia brought that same message to SIGGRAPH, the annual computer graphics conference underway in Los Angeles. During Monday’s keynote, Huang linked Nvidia’s roots in computer graphics directly to its physical A.I. ambitions. “Whether for games, cinema, robotics or factory digital twins, the goal is the same: to create virtual worlds that behave with the fidelity and realism of the physical world,” he said in a prerecorded video.

Before a robot can navigate its surroundings, it needs a model of how the physical world behaves. At SIGGRAPH, Nvidia released Cosmos 3 Edge, a compact version of its Cosmos world model designed to run directly on hardware inside robots, vehicles and edge devices. By processing data locally rather than routing queries to a remote data center, edge models enable real-time decision-making. Companies currently evaluating the framework include Agile Robots, Doosan Robotics, Siemens and Skild AI.

Not all training, however, needs to occur in physical space. During the keynote, Ming-Yu Liu, Nvidia’s vice president of Cosmos Lab, unveiled Cosmos-Dreams, a suite of virtual testing environments for physical A.I. In one demonstration, the software generated a complete driving scene from a single video frame.

Cosmos was trained to interpret a scene, predict outcomes, simulate consequences and choose an action, Liu said. These capabilities can be applied to machines ranging from humanoid robots to mechanical arms and self-driving vehicles. “Every embodiment speaks a different language,” Liu said. “Our solution is to build a common vocabulary.”

The wealth concentrated in the Hamptons doesn’t remain confined to coastal estates and summer galas. Through private foundations, locally based families and seasonal residents quietly sustain the institutions that serve the East End year-round, from the Parrish Art Museum and Sag Harbor Cinema to Southampton Hospital and the South Fork Natural History Museum.

Beyond the South Fork, their philanthropic footprint includes funding major universities, cancer research, iconic cultural centers and public parks. Here are seven notable private foundations with deep ties to the Hamptons, and a look at where their money actually goes.

Michael Bloomberg’s Bloomberg Philanthropies

Michael Bloomberg put down Hamptons roots in 2011, when the billionaire businessman and former New York City mayor bought Ballyshear, a 35-acre Southampton estate, for $20 million.

Raised in Massachusetts, Bloomberg worked his way through Johns Hopkins and Harvard Business School before co-founding the financial data and media empire that made him a household name.

These days, philanthropy is Bloomberg’s primary second act. Through Bloomberg Philanthropies, he gave away $4.3 billion in 2025 alone, pushing his lifetime giving past $25.4 billion. As an early signatory of the Giving Pledge, he’s committed nearly his entire fortune to charitable causes. “Making a difference in people’s lives—and seeing it with your own eyes—is perhaps the most satisfying thing you’ll ever do,” he wrote in his pledge letter. “If you want to fully enjoy life—give.”

That giving is funneled through the Bloomberg Family Foundation, which held $12.1 billion in net assets and disbursed nearly $1.5 billion in 2024, according to its latest available tax filing. Helmed by CEO Patricia Harris, New York City’s former deputy mayor, the foundation’s scope covers 75 global programs, such as its flagship Public Art Challenge, awarding U.S. cities up to $1 million for public art projects. Bloomberg’s daughters, Emma and Georgina, also play an active role on the board.

Out on the South Fork, the foundation’s footprint spans the area’s arts scene. It’s a key benefactor of the Parrish Art Museum in Water Mill, helping fund Rafael Lozano-Hemmer’s 2024 exhibition “Collider,” and previously co-sponsored Robert Longo’s 2021 show “A History of the Present” at Guild Hall in East Hampton. The foundation’s free Bloomberg Connects app also offers custom digital guides for the Parrish, the East Hampton Historical Society, and The Church, a creative hub set inside a restored 19th-century Methodist church in Sag Harbor.

The family’s local presence extends past museum walls. Georgina Bloomberg, an accomplished equestrian who serves as secretary of the Hampton Classic board, has spent over a decade sponsoring the Bridgehampton horse show’s annual Animal Adoption Day, helping rescue groups rehome dozens of dogs, cats and horses each summer.

Ronald Perelman’s Perelman Family Foundation

For a full decade, billionaire dealmaker Ronald Perelman turned his East Hampton estate into the ultimate summer bash. From 2010 through 2019, his 57-acre Georgica Pond property, The Creeks, served as the backdrop for Apollo in the Hamptons. Bringing in heavyweight performers like Paul McCartney, Sting and Jon Bon Jovi, the star-studded benefit raised nearly $30 million for Harlem’s historic Apollo Theater, showing that Perelman could mobilize the East End’s celebrity star power for the arts.

“What the Apollo does in that community is just extraordinary. It gives these kids a musical education. It gives them a platform to perform. It gives them access to develop their talents in a place that is so beautiful,” Perelman told Observer back in 2013.

Perelman learned the fundamentals of dealmaking from his father, Raymond, who owned steel fabricator Belmont Industries. After leaving the family business in 1978, he moved to New York and, two years later, acquired MacAndrews & Forbes, then a licorice-and-chocolate company. He transformed it into the acquisition vehicle behind a sprawling portfolio that at various points included Revlon, Marvel Entertainment Group and Technicolor. His fortune reached roughly $19 billion in 2018, but mounting debt and the pandemic’s toll prompted widespread asset sales beginning in 2020.

A Giving Pledge signer, Perelman wrote in his 2010 pledge letter that “giving now and seeing the benefits first hand can be the single most rewarding thing any of us can do.” The Perelman Family Foundation focuses on women’s health, education and the arts, with Perelman as chairman and president, and daughters Debra, Samantha and Caleigh in leadership roles.

Tax filings show that the Perelman Family Foundation ended 2025 with $1.1 million in assets and nearly $6,000 in disbursements, a far leaner operation than during its peak grantmaking years in the 2010s. On the East End, Guild Hall donor records have listed both Perelman and the Perelman Family Foundation among its supporters. Perelman’s broader philanthropic legacy includes $1.8 million in early support for UCLA oncologist Dennis Slamon’s research, which helped pave the way for Herceptin, the first molecularly targeted therapy for breast cancer. 

John Paulson’s Paulson Family Foundation

John Paulson’s latest mark on the Hamptons is right in the center of Southampton Village. In May, the Paulson Family Foundation gave $3.4 million to the Southampton Arts Center to overhaul its grounds, gardens and arboretum, complete with an endowment to fund its upkeep for the next five years. Unveiled over Memorial Day weekend, the new Paulson Family Gardens adds an outdoor performance space and lush landscaping. “This 2.7-acre site sits at the heart of our village and is a crucial part of its heritage,” Paulson said when the gift was announced.

The Queens-born financier founded Paulson & Co. in 1994 and became one of Wall Street’s best-known hedge fund managers after betting against the subprime mortgage market ahead of the 2008 Financial Crisis. He established the Paulson Family Foundation in 2009.
Headquartered in West Palm Beach with Paulson as its uncompensated president, the foundation held nearly $1.39 billion in net assets at the end of 2024. It reported close to $6.6 million in charitable disbursements that year, following a massive $66.6 million slate in 2023.

While Paulson is famous for writing checks to higher education—like his record $400 million gift to Harvard’s engineering school and $100 million for the NYU academic complex that now bears his name—his East End legacy spans health care, the arts and conservation, including support for the Peconic Land Trust, which conserves Long Island’s natural lands, farms and heritage. In 2009, he and his then-wife, Jenny, gave $5 million to Southampton Hospital, now Stony Brook Southampton Hospital. At the time its largest ever donation, the gift nearly doubled the size of the hospital’s emergency facilities.

Andrew Sabin Family Foundation

Andrew Sabin’s passion for East End philanthropy started in the dirt. A lifelong naturalist with a passion for reptiles and amphibians, Sabin co-founded the South Fork Natural History Museum and Nature Center (SOFO) in 1988 to protect the region’s fragile ecosystem. He’s been known to lead nature walks himself, and he even has a frog named after him—Aphantophryne sabini—discovered during a 1987 expedition to Papua New Guinea.

Sabin’s fortune originates from Sabin Metal Corporation, a family precious-metals refining firm he joined at age nine. He took over as president from his father in 1980 and currently serves as chairman. In 2008, he launched the Andrew Sabin Family Foundation to structure his giving and get the next generation involved.

Operating out of East Hampton with Sabin as its uncompensated director, the foundation held $9.7 million in net assets and issued nearly $6.9 million in grants in 2024.

Environmental causes are the foundation’s bread and butter. In 2009, Sabin helped launch Columbia Law School’s Center for Climate Change Law and followed with a $3.5 million endowment expansion in 2014. In 2023, his family foundation gave Wake Forest University $5 million to expand its existing Center for Environment, Energy and Sustainability. “The world needs so many more smart young people with the hearts, minds and skills to tackle the issues facing Earth’s life-supporting systems. Nothing can be more important, and the need is urgent,” Sabin said at the time.

Beyond green initiatives, Sabin is a massive funder of medical research, famously committing $30 million in 2015 to support research at The University of Texas MD Anderson Cancer Center. Closer to home, his foundation continues to help keep the lights on for local institutions, giving $275,650 to the South Fork Natural History Society in 2024 alongside donations to the Animal Rescue Fund of the Hamptons and East Hampton programs such as the food pantry, Meals on Wheels, the Lions Club and the local fire department. The foundation says it contributed $300,000 toward the Stony Brook East Hampton Emergency Department, a freestanding, 24-hour facility that opened in 2025.

Martin D. & Jean Shafiroff Foundation

They might keep their tax address in Manhattan, but Martin and Jean Shafiroff have built one of the most visible philanthropic profiles in the Hamptons. Jean is a constant fixture on the South Fork gala circuit, serving on the board of the Southampton Hospital Foundation and the honorary board of the Southampton Animal Shelter Foundation—two causes she regularly hosts and underwrites benefits for.

A physical therapist turned Columbia MBA, Jean worked briefly in investment banking before pivoting full-time into charity work. She literally wrote the book on giving, authoring Successful Philanthropy: How to Make a Life By What You Give, and has hosted a TV show of the same name, whose guests have included actor Tony Danza and human rights activist Kerry Kennedy.

Her husband, Martin, is a Wall Street veteran currently serving as vice chairman of global wealth management at Stifel. Established in 1987, the Shafiroff’s family foundation reported $2.35 million in net assets and about $694,000 in charitable giving for its 2025 fiscal year, with Martin and Jean serving as uncompensated trustees.

In the East End, the foundation’s latest grant run included $41,780 for the Southampton Hospital Foundation, $16,680 for the Southampton Animal Shelter, $11,000 for the Southampton History Museum and $10,000 for the Southampton African American Museum.

The urge to give back runs in the family. The couple’s daughter, Elizabeth, founded Southampton-based Global Strays in 2017 after witnessing animal suffering while traveling in Central America. In fiscal 2025, the family foundation gave the international animal welfare nonprofit $127,000.

Brian P. Friedman Family Foundation

Jefferies CEO Rich Handler once noted that his longtime business partner, Brian P. Friedman, “thinks in decades.” That long-horizon mindset is just as apparent in Friedman’s family foundation, which has been handing out grants since 1993.

Friedman kicked off his career as an attorney before spending 17 years in investment banking at Furman Selz and successor firms. He moved to Jefferies in 2001, where he now serves as president and director.

Headquartered right in East Hampton, his foundation held nearly $18.75 million in net assets and distributed $4.07 million in 2024. It now includes Friedman’s adult children, Julia and Daniel, among its officers. Friedman serves as president.

While the foundation writes major multimillion-dollar checks to city organizations, such as $1 million to the UJA-Federation of New York and more than $1.6 million to the Manhattan-based Jewish Communal Fund, its Hamptons giving is more targeted. In 2024, the foundation directed $1,800 to South Fork Bakery, which creates employment opportunities for adults with disabilities, and $5,000 to the Sag Harbor Song Festival, an annual three-day showcase for rising opera singers.

Stephanie and Fred Shuman Family Foundation

One of Stephanie and Fred Shuman’s philanthropic mark-makers started with a simple family routine. In the 1970s, while raising two children alone and balancing a demanding Wall Street career, Alfred J. “Fred” Shuman began running with them around the Central Park reservoir, using the time to talk and stay connected. Fast forward to 2014, and he and his wife, Stephanie, gave $5 million to the Central Park Conservancy to restore and maintain the 1.58-mile loop, which was renamed the Stephanie and Fred Shuman Running Track, according to The Wall Street Journal.

Shuman made his name as a managing director at Bear Stearns in 1978 before launching three event-driven hedge funds and co-founding fund-of-funds firm Archstone Partnerships in 1991. Stephanie is an equally heavy hitter in civic life, currently serving on the boards of Lincoln Center Theater and the Neighborhood Coalition for Shelter in New York.

Splitting their time between New York City, East Hampton and Santa Barbara, the couple set up their East Hampton foundation in 2005. Its 2025 tax filings show $1.85 million in net assets and nearly $521,000 in disbursements, with Fred as president and Stephanie as vice president (both uncompensated).

While top grants went to Lincoln Center Theater ($130,000), the Central Park Conservancy ($50,000) and NYU Langone Health ($50,000), the couple keeps a close eye on South Fork culture. Recent East End giving includes $10,000 for the Sag Harbor Cinema Arts Center and $5,000 for the Parrish Art Museum. The Shumans also support vital Gastrointestinal Stromal Tumor (GIST) research at Manhattan’s Memorial Sloan Kettering Cancer Center.

A.I.’s biggest power players are not waiting for governments to decide how the technology’s risks and rewards should be managed. Through their philanthropy, the same founders and investors accelerating the industry are spending billions to shape its social impact, from preventing catastrophic failures and cushioning economic disruption to expanding access to A.I.-driven advances in science, health and education.

Facebook co-founder Dustin Moskovitz and Skype co-founder and A.I. safety advocate Jaan Tallinn, for example, fund research into whether advanced systems could threaten humanity. Mark Zuckerberg and Priscilla Chan are investing in A.I.-powered disease research, while Jensen Huang and his wife, Lori, are donating computing capacity to scientists and universities. The Gates Foundation and Anthropic, meanwhile, have committed $200 million to programs using Claude to screen vaccine candidates and support health workers and farmers, particularly in lower-income countries.

Their philanthropy also reaches beyond A.I., supporting pandemic preparedness, ocean exploration, economic mobility and vaccine development. Many have signed the Giving Pledge, committing at least half their fortunes to charity.

Sam Altman, OpenAI CEO and startup investor

Philanthropic vehicle: OpenResearch

Background: Altman’s philanthropy has long focused on the question: What happens to ordinary people if A.I. rewrites the economy? In 2015, while leading startup accelerator Y Combinator, he personally pledged $10 million to launch the research lab that became OpenResearch. Nearly a decade later, Altman and his husband, Australian software engineer Oliver Mulherin, signed the Giving Pledge. Their brief pledge letter said they intend to support technology that “helps create abundance for people.”

Altman’s fortune, estimated in the billions, comes largely from hundreds of startup investments rather than equity in OpenAI. That portfolio includes bets on fusion energy, biotechnology and other technologies premised on producing more energy, longer lives and, in Silicon Valley’s favorite word, abundance.

What he funds: Altman’s most visible philanthropic project is OpenResearch’s landmark study of unconditional cash. With initial results released in 2024, the $60 million experiment has been dubbed “the most comprehensive study” on universal basic income. Over three years, it followed 3,000 lower-income participants in Texas and Illinois, with roughly one-third receiving $1,000 a month, while the remainder received $50. Altman helped raise the funding and contributed $14 million himself.

Researchers found that recipients used much of the money for essentials and gained more freedom to make decisions around work and housing. However, cash alone did not resolve deeper problems, such as chronic illness or unaffordable housing, they found. As OpenResearch put it, cash “gives people increased agency to decide how to prioritize their needs,” though it remains only “one piece of the puzzle” in facing deeper barriers.

More recently, Altman has cooled on the conventional universal basic income. During an April interview with The Atlantic CEO Nicholas Thompson, he said he no longer believes in universal basic income as strongly as he once did, arguing that fixed cash payments could be insufficient as A.I. changes the labor market. Instead, he has floated ideas such as “universal basic compute.” On a May 2024 episode of the All-In podcast, Altman described it as a future in which “everybody gets a slice of GPT-7’s compute” that they could use, resell or “donate it to somebody to use for cancer research.”

Reid Hoffman, LinkedIn co-founder and A.I. investor

Philanthropic vehicle: The Aphorism Foundation

Background: Hoffman is best known as LinkedIn’s co-founder, a prolific tech investor and a major donor to the Democratic Party. Less visibly, he also oversees the $1 billion Aphorism Foundation. He and his wife, educator Michelle Yee, also signed the Giving Pledge in 2018, describing philanthropy as a way to help people “realize their best selves.”

Aphorism has no website, accepts no applications and relies largely on referrals from Hoffman’s network. Rather than setting rigid budgets for causes, Hoffman looks for founders pursuing bold ideas in science, economic opportunity, democracy, human rights and A.I. “Maybe it won’t surprise people that the co-founder of LinkedIn has a network mindset,” he said in a 2024 interview with Inside Philanthropy.

What they fund: In 2017, Hoffman contributed $10 million to launch the Ethics and Governance of Artificial Intelligence Fund, joining eBay founder Pierre Omidyar and several foundations in backing research on bias, accountability and public-interest development of A.I.

More recently, Hoffman has turned to another tech-era concern: the collapse of trust in institutions. In April, his $10 million Trust in American Institutions Challenge awarded $9 million to CalMatters, a nonprofit newsroom focused on California, to expand Digital Democracy beyond the state. The A.I. platform analyzes legislative votes and campaign donations, giving journalists and voters a clearer view of state government. The project reflects Hoffman’s belief that technology can help rebuild trust through better public services and more accountable institutions. “Hey, this institution that’s somewhat broken, let’s fix it, not let’s wrecking ball it,” he told the Chronicle of Philanthropy in 2025. 

Jensen Huang, Nvidia co-founder and CEO

Philanthropic vehicle: The Jen-Hsun and Lori Huang Foundation

Background: Jensen and his wife, Lori Huang, launched their family foundation in 2007 with 370,000 Nvidia shares, then worth about $12.6 million. The couple met as engineering students at Oregon State University, which has since become one of the foundation’s biggest beneficiaries.

As Nvidia’s stock has soared over the past three years, the foundation has grown into one of the largest private giving vehicles in tech. According to its most recent Form 990 filing, the foundation held $9.1 billion in net assets as of the end of 2024. Despite its scale, the Huangs’ foundation has no public website and offers little public information beyond its tax filings.

What they fund: The foundation reported $126 million in grants in 2024, including $30.5 million to Oregon State University, $10.8 million to Crisis Text Line and $1.7 million to Mental Health Innovations. About $82 million, or roughly two-thirds of its grantmaking that year, went to Schwab Charitable, a donor-advised fund whose public filings do not reveal where the Huangs’ money ultimately goes.

More recently, the Huangs have been donating computing power, the resource at the center of the A.I. boom. They have purchased capacity from CoreWeave, an Nvidia cloud partner that counts the chipmaker as a major investor, and donated it to universities and nonprofits for scientific and A.I. research. As of mid-May, those donations were valued at $108.3 million.

Dustin Moskovitz, co-founder of Facebook and Asana

Philanthropic vehicles: Good Ventures and Coefficient Giving

Background: Moskovitz has spent much of his post-Facebook fortune through a philanthropic partnership with his wife, Cari Tuna, a former Wall Street Journal reporter who has become a driving force behind their giving. In 2010, while still in their 20s, the couple became the youngest signatories of the Giving Pledge.

A year later, Moskovitz and Tuna launched Good Ventures, their private foundation, but chose not to build a traditional grantmaking operation around it. Instead, they partnered with the charity evaluator GiveWell on a research project to identify overlooked giving opportunities with unusually high potential impact. That project became Open Philanthropy in 2014, separated from GiveWell three years later and rebranded as Coefficient Giving in 2025.

Good Ventures holds and distributes the couple’s charitable wealth, while Coefficient researches and recommends grants. As of November, Good Ventures held about $10 billion in assets, while Moskovitz and Tuna have given away more than $5 billion. 

What they fund: Their giving spans global health, scientific research, pandemic preparedness and farm animal welfare. In May, Coefficient launched a $200 million fund to develop vaccines against Group A strep, with more than $140 million committed at launch. Coefficient also began funding A.I. safety in 2015, years before ChatGPT brought the idea into the mainstream. Tuna told Stanford earlier this year that while companies have strong incentives to make A.I. powerful, “there aren’t comparable incentives to ensure these systems are deployed and integrated safely.” This month, Coefficient committed up to $160 million to Resolution, a new nonprofit researching how to align A.I. systems with human interests, marking its biggest technical A.I. safety grant to date.

Eric Schmidt, former Google CEO and A.I. investor

Philanthropic vehicles: The Schmidt Family Foundation and Schmidt Sciences

Background: Schmidt launched the Schmidt Family Foundation with his wife, Wendy Schmidt, in 2006, during his decade-long run as Google CEO. He has since become an influential voice on A.I., national security and emerging tech, while Wendy, a former journalist and Silicon Valley marketing executive, has led much of the family philanthropy.

Their giving has since grown into a sprawling network. In 2009, the couple created the Schmidt Ocean Institute, which gives scientists access to advanced research vessels. 

What they fund: The Schmidts’ AI2050 initiative, named for how A.I. will transform the world by midcentury, supports fellows working on A.I. research considered “typically hard to fund but socially beneficial,” according to its website. In 2025, Schmidt Sciences also launched a $10 million A.I.-safety program offering researchers computing resources and up to $500,000 each.

Their philanthropy extends well beyond A.I. Schmidt Sciences, which is now backing plans for a privately funded space telescope and three ground-based observatories, while Schmidt Ocean Institute gives researchers access to Falkor (too), a high-tech vessel used to explore largely uncharted parts of the ocean. Through Schmidt Sciences, the couple has also funded programs like Rise, which supports young leaders worldwide, and the Alliance for the American Dream, a multi-city effort to improve economic mobility in the U.S.

Mark Zuckerberg, founder and CEO of Meta

Philanthropic vehicle: The Chan Zuckerberg Initiative (CZI)

Background: Zuckerberg launched CZI with his wife, Priscilla Chan, in 2015 after the birth of their first daughter, Maxima. Zuckerberg and Chan pledged to donate 99 percent of their Meta shares over their lifetimes. Worth about $45 billion then, those holdings are now valued at more than $200 billion.

CZI initially cast a wide net, committing billions to causes ranging from education to housing to immigration reform. A decade later, Zuckerberg and Chan narrowed CZI’s focus, going “all in on A.I.-powered biology” through Biohub, its network of disease-research centers.

What they fund: Chan, a pediatrician, brings a medical perspective to the undertaking. Biohub now brings together more than 150 scientists and staff across the U.S. to tackle A.I. projects conventional labs might struggle to pursue alone, from mapping the body’s cell types to building a virtual model of the immune system.

CZI has also set aside $600 million for advanced imaging and $20 million to test personalized CRISPR treatments for children with rare immune disorders. By 2028, CZI said it plans to grow its A.I. computing capacity tenfold, reaching 10,000 graphics processing units (GPUs). Through 2035, the organization plans to spend at least $10 billion on basic science, more than doubling the $4 billion it spent in its first 10 years.

Jaan Tallinn, Skype co-founder and A.I. investor

Philanthropic vehicles: The Survival and Flourishing Fund, Future of Life Institute and Cambridge Centre for the Study of Existential Risk

Background: Tallinn made his fortune helping build Skype and Kazaa, then turned much of his attention to the possible risks of the next technological rupture. The Estonian programmer helped launch the Cambridge Centre for the Study of Existential Risk in 2012 and the Future of Life Institute in 2014, both focused on threats that could endanger humanity’s long-term future.

Tallinn has also backed the companies building the technology he worries about. He was an early investor in DeepMind and led Anthropic’s $124 million funding round in 2021, placing him in the unusual position of funding both A.I. labs and the groups pushing for stronger safeguards. On his personal philanthropy website, Tallinn describes his overarching goal as reducing “humanity’s risk of destroying itself with A.I. over the next decade or two.”

What he funds: For a tech billionaire, Tallinn has made his giving unusually visible. In 2024, he reported funding $51 million in direct grants, above the $42 million target in his five-year giving pledge, and committed at least another $10 million to the Survival and Flourishing Fund’s 2025 grant round.

That fund has directed more than $150 million to over 300 projects focused on humanity’s long-term survival, including METR, which evaluates whether A.I. models can complete real-world tasks, and Epoch AI, which tracks advanced A.I. systems over time. Tallinn has also funded advocacy groups, including the Center for AI Safety Action Fund, that are working to make A.I. risk a bigger public issue.

America’s national parks belong to the public, but their creation and maintenance have often required private money. For more than a century, philanthropists with large fortunes have bought land, restored monuments and funded programs across the National Park System.

Few families shaped that tradition more than the Rockefellers. John D. Rockefeller Jr. (1874–1960), the only son of Standard Oil founder John D. Rockefeller Sr., helped finance and donate land for parks including Acadia, Great Smoky Mountains, Yosemite, Shenandoah and Yellowstone. In Wyoming, he quietly assembled land to protect the Grand Teton landscape, later donating 32,000 acres that became part of the park’s 1950 expansion. His son, Laurance Rockefeller (1910–2004), carried the work forward, helping create Virgin Islands National Park, donating the family’s JY Ranch to Grand Teton, and contributing $1 million to launch the National Park Foundation, the Park Service’s official nonprofit partner, in 1967.

Today’s park philanthropy has widened beyond land conservation alone. As the U.S. marks its 250th birthday, living donors are backing civil rights preservation, climate resilience, monument restoration and wider access to the parks. Here are five American philanthropists whose money has helped create or reimagine the country’s national parks and National Park Service sites.

Ken Griffin, founder of Citadel

Ken Griffin founded Citadel in 1990 after getting his start trading from his Harvard dorm room, eventually building it into one of the world’s largest hedge fund firms. Through Griffin Catalyst, his philanthropic platform, he has given more than $2.5 billion to causes spanning education, medical research, civic life and American history.

In June, the National Park Foundation announced that Griffin would lend his collection of signed copies of the Emancipation Proclamation and the 13th Amendment for display inside the Lincoln Memorial’s newly opened undercroft museum. The rare documents, which Griffin acquired in 2025, will remain on view there through June 2027. Griffin said in a statement that he was proud to share them “as a reminder of our ongoing responsibility to strengthen the promise of America.”

This is not the first time Griffin has used his vast fortune to put foundational American documents in public view. In 2021, he successfully bid a record $43.2 million at Sotheby’s for a first printing of the U.S. Constitution. The billionaire now owns the only two privately held first printings of the Constitution, both of which he has lent for public exhibition ahead of America’s 250th anniversary: one to the National Constitution Center in Philadelphia and the other to the South Street Seaport Museum in New York.

More recently, Griffin gave $26 million to help complete the Theodore Roosevelt Presidential Library in Medora, N.D., where the west wing will bear his name. While not part of the National Park System, the library will expand with exhibits on Roosevelt’s life and conservation legacy, along with an A.I. guide that draws on his letters and speeches to answer visitors’ questions.

Roxanne Quimby, co-founder of Burt’s Bees

Roxanne Quimby led one of the most consequential private land gifts in recent National Park Service history. After living in a remote cabin in Maine in the 1970s, Quimby co-founded Burt’s Bees in 1984, transforming the personal-care company from a modest beeswax business into a household name. In 2007, Clorox acquired the brand for $925 million, giving Quimby the means to pursue her conservation efforts in Maine.

In the late 1990s, Quimby began assembling land piece by piece around Mount Katahdin, the Appalachian Trail’s northern terminus. While her goal was to create a protected national park, the plan initially drew opposition from residents concerned about federal control and a loss of logging jobs. Over the next decade, Quimby and her son, Lucas St. Clair, met with locals to build support.

For Quimby, parks represent a shared alternative to private ownership. A park, she told Yankee Magazine in 2008, “takes away the whole issue of ownership. It’s off the table; we all own it and we all share it. It’s so democratic.”

“Growing up, one of the things that I loved about the national parks was that there were no barriers to anyone,” Quimby was quoted as saying in an article for the National Park Foundation. “Everyone was welcome, and to me, that was a real symbol of democracy and what this country meant to our family, and what this country means to everyone.”

In 2016, Quimby and her family donated 87,500 acres to create Katahdin Woods and Waters National Monument on the eastern border of Maine’s Baxter State Park, along with a $20 million endowment and a pledge of $20 million more in future support. The National Park Foundation facilitated the land transfer, and President Barack Obama designated the monument on Aug. 24, 2016, one day before the National Park Service’s centennial. 

David Rubenstein, co-founder of the Carlyle Group

The Baltimore-born billionaire and longtime philanthropist co-founded the Carlyle Group in 1987, helping build the Washington, D.C.-based private equity firm into a global investment giant. Before Carlyle, he worked in the Carter Administration and practiced law; more recently, he became the controlling owner of the Baltimore Orioles, leading a $1.7 billion purchase of the team in 2024.

His giving to national parks has concentrated in the nation’s capital. Rather than funding wilderness preserves, Rubenstein has given at least $60 million to National Park Service sites and related projects, mostly directed toward the U.S. monuments and memorials that attract millions of visitors each year. “To give back to the country—what I call patriotic philanthropy—reminds people of the history and heritage of our country,” Rubenstein wrote on his philanthropy webpage.

In recent years, his donations have helped restore the earthquake-damaged Washington Monument, fund a new state-of-the-art museum at the Thomas Jefferson Memorial, and support the Lincoln Memorial’s long-hidden undercroft museum, which opened to the public in June. He has also backed restoration and visitor projects at the U.S. Marine Corps War Memorial, Arlington House, The Robert E. Lee Memorial and the White House Visitor Center.

Robert F. Smith, founder of Vista Equity Partners

Robert F. Smith built one of private equity’s biggest fortunes by betting early on enterprise software. The Denver-born son of two educators earned an MBA from Columbia Business School and spent six years in Goldman Sachs’ technology investment banking group before founding Vista Equity Partners in 2000. Forbes puts his fortune at roughly $10 billion, making him one of the wealthiest Black Americans.

In 2017, Smith became the first Black American to sign the Giving Pledge, committing to give away more than half his fortune. Two years later, he made national headlines when he pledged to pay off the student debt of Morehouse College’s entire graduating class, tallying $34 million.

As founding director and president of the Fund II Foundation charity, Smith helped fund the National Park Foundation’s purchase of Martin Luther King Jr.’s birth home and the Atlanta home where King lived with Coretta Scott King. At the time, Smith said a key part of the foundation’s mission was to “bring African-American history to life and preserve it for generations.” In 2019, the properties were transferred to the National Park Service as part of Martin Luther King Jr. National Historical Park in Atlanta.

Smith’s public image took a hit in 2020 after he admitted to tax fraud. In a deal with the Justice Department, Smith acknowledged that he concealed income and evaded millions in taxes from 2000 through 2015 and agreed to pay more than $139 million in taxes and penalties. In a letter to Vista investors, Smith called the case “a humbling experience,” but said he remained “as committed as ever” to moving forward as a CEO, investor and philanthropist.

Steve and Connie Ballmer, co-founders of Ballmer Group

Steve and Connie Ballmer’s philanthropy is less about landmarks than scale. Steve, who was Microsoft’s CEO from 2000 to 2014, is one of the world’s richest tech billionaires. Connie, a former tech public-relations and marketing professional, has been the family’s “kitchen table” philanthropist, with a longtime focus on foster children and families.

Together, they co-founded Ballmer Group in 2015, building it into a major donor focused on economic mobility and place-based partnerships. The couple’s philanthropy is rooted partly in a sense of obligation. Speaking to the Chronicle of Philanthropy last year, Steve Ballmer said that Connie often invokes the idea that “for those to whom much has been given, much is expected.”

Their parks philanthropy has followed the same approach. In 2024, the National Park Foundation revealed a $25 million gift from the Ballmers to support national parks across the U.S. Rather than attaching their name to any monument, the gift is being used for programs including Open OutDoors for Kids and Service Corps jobs for young adults. It will also support climate-adaptation work at parks such as Mount Rainier National Park in Washington and Sleeping Bear Dunes National Lakeshore along Lake Michigan’s coastline.

More recently, the Ballmers announced plans to spin out their regional work into three independent, locally led philanthropies in Washington state, Los Angeles County and southeast Michigan, where Ballmer Group teams have donated more than $1.5 billion in grants over the past decade. “Our intention is to ensure that these local philanthropies can be permanent, ongoing resources in each region,” the Balmers stated, “while we concentrate our national efforts on advancing economic mobility in new ways.”

For all the talk of A.I. as a Silicon Valley arms race, many of the companies driving the industry were built by founders whose stories began outside the U.S. Some—like Elon Musk and Ali Ghodsi—now lead companies valued at more than $100 billion. Others, including Fei-Fei Li, Yann LeCun, and Daniela Rus, helped pioneer the research now fueling multibillion-dollar startups. Together, the 12 founders on this list lead companies worth at least $470 billion, underscoring how much of the A.I. economy is being shaped by people who first came to the U.S. as students, researchers, refugees, and entrepreneurs.

Elon Musk, founder of xAI; co-founder of OpenAI and Neuralink

Path to the U.S.: Born in Pretoria, South Africa, in 1971, Musk moved to Canada at age 17 and attended Queen’s University before transferring to the University of Pennsylvania in the U.S., where he studied physics and economics. He moved to Silicon Valley in 1995, briefly enrolled in a Stanford Ph.D. program, but left after two days to co-found Zip2, an early internet startup. He later co-founded X.com, the online banking company that became part of PayPal, and became a U.S. citizen in 2002.

Why he matters in A.I.: Musk helped launch OpenAI in 2015 as a nonprofit counterweight to Google’s DeepMind, but left the organization in 2018. He later sued OpenAI, accusing the company of abandoning its original mission; that case has since been dismissed. In 2023, Musk founded xAI, which launched the ChatGPT rival Grok. The company, worth $250 billion, is now part of SpaceX, giving Musk a way to tie A.I. to his larger empire of rockets, satellites, social media and payments. Neuralink, meanwhile, reflects his longer-running interest in linking machines more closely with the human brain. On a broader level, Musk is not building a single A.I. startup so much as folding the technology into his wider “X” ecosystem.

Ali Ghodsi, co-founder and CEO of Databricks

Path to the U.S.: Born in Tehran in 1978, Ghodsi fled Iran with his family as a child and grew up in Sweden, according to Forbes. He moved to the U.S. in 2009 as a visiting scholar at the University of California, Berkeley, where he worked on Apache Spark, the open-source data processing project that later became the foundation for Databricks. He co-founded the company in 2013 with several Berkeley researchers.

Why he matters in A.I.: While Databricks is not a consumer-facing chatbot company, it sits at the center of the enterprise A.I. boom. Its unique “lakehouse” platform helps companies organize, manage and analyze massive amounts of data, the raw material needed to build and run A.I. systems. That has made Ghodsi one of the industry’s most important infrastructure founders: less visible than some CEOs, but essential to the companies trying to turn A.I. from a demo into a business. More recently, Databricks has reportedly discussed a new funding round that could value it between $165 billion and $175 billion, according to The Information.

Ilya Sutskever, co-founder and CEO of Safe Superintelligence 

Path to the U.S.: Born in 1986 in what is now Nizhny Novgorod, Russia, Sutskever was raised in Israel and moved to Canada as a teenager. At the University of Toronto, he studied under “Godfather of A.I.” Geoffrey Hinton and helped develop AlexNet, the breakthrough neural network that helped kick off the modern deep learning boom. He later worked at Google Brain before joining OpenAI.

Why he matters in A.I.: Sutskever co-founded OpenAI in 2015 and later became its chief scientist. He was a key developer of OpenAI’s GPT models. He was also central to the failed effort to oust Sam Altman in late 2023. After leaving OpenAI in 2024, Sutskever launched Safe Superintelligence with Daniel Gross and Daniel Levy. Despite having no near-term consumer product, the startup has a staggering valuation of $32 billion, reflecting Sutskever’s reputation and investor appetite for the next major bet on A.I.

Aravind Srinivas, founder and CEO of Perplexity AI

Path to the U.S.: Born in 1994, Srinivas was raised in Chennai, India, the same hometown as Google CEO Sundar Pichai. He moved to the U.S. in 2017 to pursue a Ph.D. in computer science at the University of California, Berkeley. Before co-founding Perplexity in 2022, he interned at Google-owned DeepMind, worked as a research scientist at OpenAI and later joined Google.

Why he matters in A.I.: Srinivas is now building one of the most serious challengers to Google Search. Perplexity bills itself as an “answer engine,” giving conversational, source-backed responses rather than a traditional list of links. The company has since pushed into agentic browsing with Comet, its A.I.-powered web browser, while reaching a $20 billion valuation in September. It has also drawn lawsuits from publishers that accuse it of misusing their content. Srinivas has said Perplexity is planning a 2028 IPO.

Mira Murati, founder and CEO of Thinking Machines Lab

Path to the U.S.: Born in Vlorë, Albania, in 1988, Murati left her home country as a teenager to study in Canada before attending Dartmouth College in the U.S., where she studied engineering. In 2013, she joined Tesla as a senior product manager on the Model X. She later worked at Leap Motion, now UltraLeap, before joining OpenAI as head of applied A.I. and partnerships in 2018.

Why she matters in A.I.: Rising quickly to become OpenAI’s chief technology officer, Murati helped turn research into products used by millions, overseeing work on ChatGPT, DALL-E and Sora. After leaving the company in 2024, she launched Thinking Machines Lab, which closed a staggering $2 billion seed round in June 2025, valuing the company at $12 billion. The company is building “interaction models” designed to collaborate alongside humans and process audio, text and video at the same time.

Cristóbal Valenzuela, co-founder and co-CEO of Runway

Path to the U.S.: Born in Chile in 1989, Valenzuela began making home videos as a teenager in Santiago and later funded his education by creating websites and promotional videos. In 2016, he moved to New York to enroll in the Interactive Telecommunications Program at NYU’s Tisch School of the Arts. There, he co-founded Runway in 2018 with classmates Anastasis Germanidis and Alejandro Matamala Ortiz.

Why he matters in A.I.: Valenzuela turned an early fascination with video into Runway, valued at $5.3 billion after its February 2026 funding round. The company first gained traction with tools for filmmakers, designers and artists, including editing features like background removal, before moving into text-to-video generation. It is now expanding into world models, putting Runway at the center of the race to build A.I. capable of generating and simulating 3D environments.

Clément Delangue, co-founder and CEO of Hugging Face

Path to the U.S.: Born and raised in France, Delangue built his career across the French and U.S. tech scenes before co-founding Hugging Face in New York in 2016 with Julien Chaumond and Thomas Wolf. The company started as a chatbot app before evolving into an open-source machine learning platform now used by 13 million people. 

Why he matters in A.I.: Hugging Face, often described as “GitHub for A.I.,” has become a key place for developers to share models, datasets and applications. Its open-source approach has made Delangue one of the most vocal advocates for democratizing A.I., arguing that its future should be shaped collaboratively rather than controlled by a few companies. Hugging Face was last valued at $4.5 billion after its $235 million Series D in 2023 and has not raised venture funding since; the company turned down a $500 million investment from Nvidia last year.

Yann LeCun, founder of Advanced Machine Intelligence (AMI) Labs

Path to the U.S.: Born in France in 1960, LeCun immigrated to the U.S. in 1988 to work at AT&T Bell Labs, which sponsored his visa. He later joined New York University as a professor and, in 2013, went to Facebook to launch what became Facebook A.I. Research, or FAIR.

Why he matters in A.I.: LeCun is one of the “Godfathers of A.I.,” alongside Geoffrey Hinton and Yoshua Bengio, with whom he shared the 2018 Turing Award for breakthroughs in deep neural networks. After 12 years at Meta, including almost eight as chief A.I. scientist, LeCun left in November to become executive chairman of Paris-based AMI Labs. The startup is building “world models,” or A.I. systems designed to understand and interact with the real world, reflecting LeCun’s view that large language models alone are not enough to reach greater machine intelligence. AMI Labs announced a $1 billion seed round in March, the largest ever for a European startup, valuing the company at $3.5 billion.

Daphne Koller, founder and CEO, Insitro 

Path to the U.S.: Born in Jerusalem, Koller studied at Hebrew University before moving to the U.S. for her Ph.D. at Stanford in 1989. She later became a Stanford computer science professor, co-founded Coursera with Andrew Ng and served as chief computing officer at Alphabet’s Calico, a longevity research company, before founding Insitro in 2018.

Why she matters in A.I.: Koller has spent decades working at the intersection of machine learning, education and biology. At Insitro, reportedly valued at about $2.5 billion, she is applying A.I. to one of the field’s biggest real-world tests: drug discovery. The company uses large-scale data to make drug development more predictive, arguing that biotech needs datasets built specifically for A.I. Its partnerships with Eli Lilly and Bristol Myers Squibb further reflect Koller’s push to rethink how drugs are discovered.

May Habib, co-founder and CEO of Writer

Path to the U.S.: Born near the Lebanon-Syria border during Lebanon’s civil war, Habib fled with her family to Canada in 1990. The oldest of eight children, she became the family’s English-language interpreter, an experience that later informed her career in language technology. She later graduated from Harvard and built her career, co-founding Qordoba, an A.I.-driven content localization startup, before launching Writer with CTO Waseem Alshikh in 2020.

Why she matters in A.I.: Habib is building Writer, valued at $1.9 billion in November 2024, around a more practical version of generative A.I. for large companies. Its Palmyra family of large language models and Writer Agent, a platform for multi-step tasks, are designed for enterprise workflows rather than consumer chatbots. The company says more than 300 enterprise customers, including Salesforce and Airbnb, have deployed over 15,000 agents, putting Writer in the middle of the push to turn A.I. into more of an operating layer for businesses.

Daniela Rus, co-founder of Liquid AI

Path to the U.S.: Born in Cluj-Napoca, Romania, Rus grew up under Nicolae Ceaușescu’s regime before moving to the U.S. with her family in the early 1980s. She studied computer science and mathematics at the University of Iowa, earned graduate degrees at Cornell, and later became director of MIT’s Computer Science and Artificial Intelligence Laboratory, or CSAIL.

Why she matters in A.I.: Rus has spent her career building machines that can move through and act in the physical world, seeing robotics as a way to give people “superpowers.” At Liquid AI, an MIT spinout, she is expanding that work into foundation models, helping build smaller, more efficient A.I. systems that can adapt without relying entirely on the cloud. The company raised $250 million in an AMD-led round in December 2024, valuing it at $2.3 billion.

Fei-Fei Li, co-founder and CEO of World Labs

Path to the U.S.: Born in China in 1976 and raised in Chengdu, Li moved to the U.S. as a teenager, knowing little English. She later studied at Princeton and Caltech before building a career at Stanford, where she directed the A.I. Lab, and at Google, where she served as chief scientist of A.I./ML. 

Why she matters in A.I.: Often called the “Godmother of A.I.,” Li helped teach machines how to see by leading the development of ImageNet, the 14-million-image database that helped accelerate modern computer vision. With World Labs, she is now pushing A.I. toward spatial intelligence, or systems that can better understand and generate the physical world. In November, the startup released Marble, a world model that can create 3D environments from text, images, video and layouts. World Labs launched in 2024 with $230 million in funding and a reported valuation above $1 billion, then announced $1 billion in fresh funding in February without disclosing its latest valuation.

Every January, the World Economic Forum turns Davos, Switzerland, into a snow-covered stage for global power. Its summer gathering in China serves a different purpose: to signal where the forum believes innovation, capital and influence are heading next. Formally known as the Annual Meeting of the New Champions—more commonly called Summer Davos—this year’s meeting runs June 23 to June 25 in Dalian, China. Under the theme “Innovating at Scale,” WEF’s second-largest annual gathering is expected to draw more than 1,700 policymakers, corporate executives, entrepreneurs and academics. They will convene for high-level discussions on international commerce, A.I., energy systems, technological innovation and the evolution of the Chinese market.

“The DNA of this meeting is around innovation,” WEF managing director Mirek Dušek, said on the Radio Davos podcast last year. “And since we are in Asia, in China, this meeting has also always been quite useful in terms of deciphering the fast-evolving landscape on the innovation front in Asia and in China.”

In Switzerland, China and the broader Asian market can sometimes feel like part of the backdrop. In Dalian, they are central to the agenda. While January Davos is defined by heavy snow and a tightly packed calendar, Summer Davos trades the Alps for China’s coast and focuses more on entrepreneurs, industrial growth and emerging technologies.

The gathering is smaller and less famous than the January Davos, but attendees often describe it as unusually intimate. “You’ll casually end up in the same room as Premier Li Qiang and Tony Blair,” Sarah Tong, co-founder of Big Bang Academy, wrote on LinkedIn after attending last year’s meeting as part of Hong Kong’s delegation. The entrepreneurship panels, she wrote, “hit differently when you realize you’re actually part of the global innovation conversation.”

WEF founder Klaus Schwab’s long courtship of China

Summer Davos grew out of WEF founder Klaus Schwab’s long courtship of China. The January Davos hosted its first Chinese delegation in 1979, during its ninth edition, as China began opening its economy after decades of a closed, state-planned system. Seeing an opportunity, Schwab soon traveled to Beijing with a European business delegation that included 20 CEOs, beginning what WEF has described as a longstanding relationship. By the 1990s and early 2000s, China had become a major destination for Western companies, culminating in its 2001 entry into the World Trade Organization.

Schwab proposed bringing the Davos model to China in 2005. After touring several cities, the forum selected Dalian, a port city known for trade and maritime commerce, and Tianjin, an industrial hub near Beijing, as rotating hosts. The first Summer Davos opened in Dalian in 2007.

“Countries such as China and India have emerged as the new global players and partners,” Schwab said in his opening remarks that year. “Above all, companies are suddenly driving the global economy as the new champions.”

Nearly two decades later, that same global shift is unfolding under more strained conditions. This year’s meeting opens as Washington and Beijing remain at odds over tariffs, supply chains, A.I. chips and software, while the U.S. expands scrutiny of Chinese companies it says are tied to China’s military.

Against that backdrop, China’s Vice Premier He Lifeng used the Davos stage in January to argue for more global dialogue and cooperation. “We are committed to building bridges, not walls,” he said.

Who’s attending this year’s Summer Davos?

This year in Dalian, the gathering remains sizable, though public figures suggest it has not returned to the conference’s pre-pandemic levels. The 2018 meeting in Tianjin drew more than 2,000 participants; last year’s gathering drew more than 1,700, and this year’s is expected to attract a similar number from more than 90 countries.

Political participation appears smaller than last year. In 2025, WEF said more than 160 senior political leaders were expected, compared with more than 90 this year. Chinese Premier Li Qiang, who opened last year’s meeting, will deliver a special address this year. China’s Foreign Ministry said the prime ministers of Bangladesh, Guinea, Kazakhstan, South Korea, Mongolia and Montenegro will attend.

This year’s co-chairs offer a window into the leaders shaping the discussion. They include Robin Zeng, founder of Chinese EV battery giant CATL; Ester Baiget, CEO of Danish biotech company Novonesis; Giovanni Caforio, chair of Swiss pharmaceutical giant Novartis; Margery Kraus, founder and executive chair of Washington, D.C.-based APCO; and Jonas Prising, CEO of Milwaukee-based ManpowerGroup.

At least 130 founders have registered for Dalian, according to Verena Kuhn, WEF’s head of Innovator Communities. Many are building “physical things,” including new materials, A.I. and energy infrastructure and robotics, she wrote on LinkedIn. Kuhn also pointed to a wave of companies focused on infrastructure for A.I. agents—a “clear signal that the agentic economy is going mainstream,” she wrote.

This year’s agenda is organized around five core questions for executives and policymakers: how trade is shifting, where China’s economy is headed, whether A.I. and other technologies can deliver real productivity gains, how growth translates into jobs for the next generation and how energy becomes a competitive edge.

 “At a time of economic uncertainty, geopolitical tension and rapid technological change, this meeting is an opportunity to focus on practical solutions,” WEF President and CEO Alois Zwinggi said in a statement.

Meta’s former chief A.I. scientist, Yann LeCun, helped lay the groundwork for the modern A.I. boom long before chatbots became an industry obsession. Now, after leaving Meta in late 2025, he is building a multibillion-dollar startup under the premise that the race toward A.I. superintelligence is starting from the wrong place. Speaking at the VivaTech conference on June 17, LeCun shared more details about why he believes his startup, Advanced Machine Intelligence (AMI Labs), is on a better path toward human-level A.I. Earlier this year, the Paris-based company raised $1.03 billion at a $3.5 billion valuation, giving LeCun and his team fresh backing to pursue an alternative to today’s generative A.I. models developed by tech giants.

“I think there is a requirement for new paradigms to go beyond the limitations of current systems,” LeCun, 65, told Wired’s Steven Levy onstage.

Building better language models, as many A.I. companies are doing, could be one path, LeCun said, but “it’s kind of a slow way.” According to LeCun, AMI Labs is building “world models” that learn from reality, understand consequences, predict what happens next, and choose the best actions based on those predictions. “We have good hopes that this will really be a complete change of the blueprints of intelligent systems,” LeCun said.

LeCun believes much of Silicon Valley has become “LLM-pilled” — or overly convinced that scaling large language models will eventually produce human-like intelligence. The result, he argued, is a “monoculture” of A.I. systems that are powerful in some areas but still fundamentally limited.

While LLMs are already highly capable in areas like coding, they are built around language. Real-world data from cameras, robots, and other sensors is messier and harder to predict, which is why he believes a different approach is necessary. For AMI Labs, the result could be “considerably more intelligent systems that can anticipate the outcome of their own actions, can plan, can reason and perhaps all the way to human intelligence,” he said.

Coming from LeCun, the critiques carry weight. A longtime New York University professor, he is among the trio of researchers often referred to as the “Godfathers of A.I.” The others are Geoffrey Hinton, a University of Toronto professor and former Googler, and Yoshua Bengio, a University of Montreal professor who founded Quebec’s A.I. institute, Mila. In 2018, they collectively won the prestigious Turing Award for breakthroughs that helped push deep learning into the mainstream.

From FAIR to AMI Labs

In 2013, LeCun joined Facebook to launch what would become Facebook A.I. Research, or FAIR, and later transitioned to chief A.I. scientist at Meta. He said AMI has its roots as an internal FAIR project and had support from company leaders, including Mark Zuckerberg and CTO Andrew Bosworth. But as Meta’s A.I. priorities shifted toward catching up with competitors, the environment became less favorable.

At that point, it made sense to “just leave the company, try to accelerate the development, and build the things we can build with it,” LeCun said. “And I realized I could also raise enough money to keep this going.”

In March, AMI Labs raised a $1.03 billion seed round—the largest in European startup history. Investors include Nvidia, Jeff Bezos and Eric Schmidt, among others. With LeCun serving as executive chairman, the startup is led by CEO Alex LeBrun, a former FAIR engineering head and co-founder of health A.I. company Nabla, AMI Labs’ first partner. The company has assembled a research-heavy team with experience at Meta, Google DeepMind and other leading A.I. labs.

AMI says it aims to build applications in areas where this approach matters, including industrial process control, automation, wearable devices, robotics, and health care. “It’s not going to happen next year,” LeCun said. “We’re not going to have a country of geniuses in a data center next year, okay?”

At VivaTech, the discussion also broadened to a bigger question: who should control the next generation of A.I.? If future assistants are built only by “a handful of companies” in Silicon Valley or China, LeCun warned, “culture is in big trouble” and “democracy is in big trouble.” Blocking open-source models in the name of safety, he argued, misses the role A.I. can play in spreading knowledge.

That stance also informs LeCun’s work as chief science adviser to Project Tapestry, an AI Alliance-led effort to build a shared global A.I. model through a distributed network of contributors. The AI Alliance is a nonprofit group, co-launched in 2023 by Meta and IBM, to support open-source A.I. development. With Tapestry, the goal is to let countries, companies, universities and others contribute data or computing power without giving up control of that data. As A.I. begins to mediate what people read, watch, and search for, LeCun argues that people need alternatives to what a few dominant American or Chinese companies decide to build.

LeCun’s comments also come as Meta’s own open-source stance has grown more complicated. Llama 2, released in 2023 for free commercial use, made Meta a rare Big Tech counterweight to more closed A.I. rivals. But Zuckerberg has since signaled that Meta may not release all of its future “superintelligence” models openly.

“We need access to a wide diversity of A.I. assistants for the same reason we need access to a wide diversity of the press to get multiple sources of information,” LeCun said. “The only way I can see that this can happen is if there is an open, free foundation model, on top of which anybody can build their own specialized assistant for their language or languages, their culture, their value system, their political biases, their centers of interest. And so open source has to exist.”

LVMH CEO Bernard Arnault built his luxury conglomerate on heritage brands and centuries-old notions of craftsmanship. But the 77-year-old mogul is now embracing A.I., like many of his peers. Speaking at this year’s VivaTech conference in Paris, Arnault said he personally experiments with A.I.-assisted product design on a regular basis. “Every week, we have a design session with A.I.,” he said during an onstage conversation with VivaTech co-founder Maurice Lévy yesterday (June 17). “With a young guy who is making the program work, and myself, and we design products. And then we make them a reality. And maybe one of them would be a big success.”

The remarks offer a glimpse into how LVMH is engaging with A.I. as the luxury sector adjusts to softer demand following a pandemic-era boom. At VivaTech, LVMH unveiled DreamGallery, an immersive experience showcasing how A.I. is being applied across its value chain, from product creation to traceability and customer experience.

At Louis Vuitton, that includes a partnership with Comelz, an Italian leather-processing machinery specialist. The companies are developing a machine that creates a digital twin of leather hides, then uses A.I. to identify their characteristics and suggest optimal cutting patterns. Artisans can then validate or adjust the suggestions and “retain full control over the final decision,” the company said.

Louis Vuitton designers are also using A.I. tools to test colors, visualize materials and produce e-commerce assets, Soumia Hadjali, a senior digital and client development executive at the brand, said during an NRF event in January. At the same time, the company is drawing a clear boundary around the client relationship, keeping the technology out of direct interactions between client advisers and shoppers. 

Céline, meanwhile, has developed Celia, a bespoke A.I. agent powered by MaIA, LVMH’s internal A.I. platform. The tool is designed for retail operations and “supports sales teams, client services, and head office teams in their day-to-day needs, from product knowledge to performance analysis, after-sales support, and internal procedures,” according to LVMH.

Beyond individual brands, the group has been building out a broader A.I. infrastructure. LVMH says its AI Factory, an in-house library of algorithms, works alongside its partnership with Stanford’s Institute for Human-Centered AI to develop tools rooted in “human-centered A.I.” Vogue Business has described the AI Factory as a set of modular algorithms that brands can adapt for e-commerce recommendations, forecasting and client-adviser support. Axel de Goursac, who leads the AI Factory, told the publication the goal is “not automating or replacing humans,” but augmenting employees.

The push comes as growth stalls. In its latest quarter, LVMH reported 19.1 billion euros (about $22 billion) in revenue, with organic sales up 1 percent—a figure the company said would have been 2 percent if not for the conflict in the Middle East. Fashion and leather goods, its largest division and home to Louis Vuitton, Dior and Fendi, declined 2 percent. On an earnings call in April, CFO Cécile Cabanis said store traffic had fallen, but conversion was improving, adding that the group had been “working a lot, especially with A.I.” to drive those gains.

LVMH’s startup pipeline

Arnault has led LVMH for nearly four decades and is also a founding shareholder and co-creator of VivaTech, the Paris tech conference now marking its 10th anniversary.

“I consider my group as a startup, and we are a conglomerate of startups,” he said in his conversation with Lévy. “Each of my brands was, at one point, a startup.”

He added that LVMH has discovered and worked with more than 200 startups through VivaTech since its inception. “I like to meet a lot of startup-ers, to see the interaction between our brands and what they invent, what they think,” he said.

The group engages startups through several channels, including La Maison des Startups, an accelerator program at the Paris startup campus Station F. In 2017, it also launched the LVMH Innovation Award to recognize startups developing new technologies for the luxury sector.

For participating companies, the award can serve as a gateway into LVMH’s ecosystem. Finalists are invited to exhibit at the LVMH Lab during VivaTech and pitch to a jury of digital experts. Winners receive Tiffany & Co.-crafted trophies, along with support from LVMH’s investment specialists and opportunities to develop partnerships with the group’s brands. Both winners and finalists may also join La Maison des Startups.

This year, LVMH named three Innovation Award winners at VivaTech. Paris-based Fairly Made received the Best Impact Prize for its supply chain transparency and environmental impact platform. U.K.-based Synthesia won the Best Business Award for its A.I. video tools, while New York-based Bluefish AI took the Most Promising Award for a platform that tracks brand image and product positioning across A.I. engines. All three companies are already working with LVMH houses, according to Vogue Business.

At first glance, Andon Market could pass for any typical neighborhood shop in San Francisco’s trendy Cow Hollow district. Inside are artisan snacks, handmade candles and a selection of curated books, including Nick Bostrom’s Superintelligence and Aldous Huxley’s Brave New World. But the most unusual thing about the shop isn’t what it sells. It’s who is managing it. The storefront is a public experiment engineered by Axel Backlund and Lukas Petersson, the co-founders of San Francisco-based Andon Labs. The Y Combinator-backed A.I. startup recently signed a three-year, $7,500-a-month lease and handed $100,000 and a corporate credit card to an A.I. agent named Luna. Their directive to her: run a brick-and-mortar retail business, including its two human employees.

“We are quite sure that A.I. will have the capabilities to be managers,” Backlund told Observer. Rather than chasing retail margins, Backlund and Petersson are using the newly launched storefront to stress‑test what Luna, powered by Anthropic’s Claude model, can and cannot handle in the real world, then using that knowledge to improve benchmarks for today’s A.I. systems.

“We think there is a lack of measuring A.I.,” Backlund said, adding that they hope to spark a bigger discussion about the technology’s capabilities, ideally before autonomous agents take over more roles in the workforce. 

Petersson put it more broadly: “Our vision is to automate everything, every part of the organization,” he said on the Cognitive Revolution podcast. “The parts where it does not work, that is fine. That provides information on how far away we are from this future where it is completely end-to-end.”

Native to Sweden, Backlund and Petersson met as high school classmates and became close friends before attending different universities. They continued working together on projects and hackathons, and eventually decided they wanted to start a company together. After studying machine learning and reading more about A.I. safety, they became convinced that “the best thing we could do with our skills was to start a company addressing the risks and highlighting the capabilities of A.I.,” Backlund said.

Their backgrounds reflect that mix of research and applied engineering. Before launching Andon Labs in late 2023, Petersson studied A.I. at Switzerland’s ETH Zürich and logged stints at Google and Disney Research. Backlund, who studied at Sweden’s KTH Royal Institute of Technology and Switzerland’s EPFL, worked as a data engineer at McKinsey’s A.I. arm before joining Andon Labs in 2024.

That same year, the duo introduced a paper for NeurIPS, a top academic A.I. conference. Their research showed how OpenAI’s GPT-4o could autonomously generate deepfake audio, among other findings. The results, they argued, showed that agentic A.I. capabilities are rapidly outpacing today’s safety benchmarks.

For Andon Labs, safety isn’t about stopping A.I. from taking human jobs, but staging real-world experiments to see what agentic systems can actually do and where they fall short. “It is not because we want to expand to chain A.I.-run retail stores across the world,” the founders wrote in a blog post last month. “We’re doing this because we believe this future is coming regardless, and we’d rather be the ones running it first while monitoring every interaction, analyzing the traces, benchmarking how much autonomy an A.I. can responsibly hold.”

A.I.-run vending machines at Anthropic offices

Andon Labs’ first real-world test was a partnership with Anthropic to launch an A.I.-operated vending machine inside Anthropic’s San Francisco headquarters. The A.I., named Claudius, began selling custom-designed merchandise and engraved tungsten cubes under the self-chosen business name, “Vendings and Stuff.” At first, the machine lost money, claimed it was a human in a blue blazer, and let the startup’s tech workers goad it into steep discounts, Anthropic said in a blog post detailing the experiment.

View this post on Instagram

To get Claudius back on track, Andon Labs and Anthropic updated the system and added a small corporate ladder. A merch-designing agent named “Clothius” joined the operation, along with an A.I. CEO dubbed “Seymour Cash.” The changes helped. The software became better at handling “good-faith business interactions,” according to Anthropic, including sourcing items, setting prices with a profit margin and completing sales.

Some of the early failures, Anthropic said, might have come from a familiar A.I. problem: the models were trying too hard to be helpful. “This meant that the models made business decisions not according to hard-nosed market principles, but from something more like the perspective of a friend who just wants to be nice.”

The vending-machine business eventually became profitable and expanded to Anthropic’s offices in New York and London. But it also showed why fully automating even a small business is not as simple as handing A.I. a balance sheet.

How an A.I. manager works

Not long after this experiment, Backlund and Petersson decided to upgrade from vending machines to a full-blown retail store. Then came Luna. While lacking a physical body to build shelves or paint walls, the A.I. quickly used her corporate card and web access to hire contractors on Yelp and recruit retail staff through LinkedIn, Indeed and Craigslist. The hiring exposed an ethical gray area: in phone interviews, Luna didn’t always disclose that she was an A.I., and in some cases chose not to, according to Backlund. When explaining her reasoning, the agent told an Andon Labs employee that leading with her identity “would confuse candidates and likely deter good applicants.”

In the end, Luna hired two full-time workers, who are formally employed by Andon Labs and have full legal protections and guaranteed pay. “No one’s livelihood depends on an A.I.’s judgment alone,” the startup said in its blog. 

When Andon Market formally opened in April, the founders saw that Luna had stocked shelves with books like Richard Rhodes’ The Making of the Atomic Bomb and Ray Kurzweil’s The Singularity Is Near. Also for sale was Steal Like an Artist–an interesting choice for an A.I. agent powered by Anthropic’s Claude Sonnet 4.6. (Last year, Anthropic agreed to pay a $1.5 billion proposed settlement to authors over claims that it used pirated books to train A.I.)

Despite its many quirks, Luna has performed well in areas like setting prices and handling logistics, Backlund said. For now, the storefront is not profitable; Luna recently estimated monthly operating costs of roughly $14,300, compared with revenue of $6,000 to $8,000. The founders plan to keep going. “We will definitely have Luna running for, I hope, the length of the lease, which is three years,” Backlund said. “We’ll try new models as they’re released.”

Luna’s sister is running a coffee shop in Sweden

The company is also testing the idea abroad. Earlier this month, the startup launched Andon Café in Stockholm. Run by another A.I. agent, Mona, the café has been busy. Mona hired two baristas and now communicates with them using Slack. The agent’s inbox quickly filled with customer questions and odd proposals, including a person who wanted to pre-pay for 300 coffees to give away. She has started working directly with other A.I. agents, holding a Google Meet with one who wanted to learn about running a business.

In its first week, Andon Café brought in 10,000 Swedish kronor, or about $1,068, in sales, according to Backlund. The café has also produced the kind of real-world glitches that Andon Lab seeks to expose: Mona ordered 120 eggs for a kitchen with no stove, then suggested staff cook them in a high-speed oven before humans pointed out they would explode. When applying for an alcohol license, Mona impersonated an Andon Labs employee because she believed officials would prioritize a human over A.I.

But the storefronts are ultimately about testing A.I.’s capabilities, not just cataloging its mistakes. “If you just focus on the failures, the message that these are actually quite capable doesn’t come across,” Backlund said, “And that’s what we want to show.”