Here Are Some Of The Biggest Winners Of This Year’s AI Infrastructure Boom

It’s been a big year for AI infrastructure deals. In just the last month, OpenAI, Oracle, Nvidia, AMD and others have announced epic transactions worth hundreds of billions of dollars. The deals are complex, unconventional—and often circular: Nvidia’s up to $100 billion investment in OpenAI, announced late last month, for instance, will allow OpenAI to purchase Nvidia GPUs to build out its own data center capacity. In a similar deal that flips who gets equity, AMD and OpenAI announced their own strategic partnership on Monday in which OpenAI got a warrant for up to 160 million shares of AMD common stock, or 10% of AMD’s shares, that vests in tranches based on OpenAI using 6 gigawatts of AMDs GPUs over a period of time.

Behind this spate of deals is something of a stampede mentality. “The world needs much more compute,” OpenAI CEO Sam Altman posted on X announcing the AMD deal. Altman has been trumpeting that “more compute is more important now than ever before to succeed at our mission” for years.

“There’s so much impatience and desire to move quickly and fear of getting left behind that there’s a very high premium on getting the most you can the fastest,” says Stella Biderman, executive director of AI nonprofit EleutherAI, which trained an open-source version of GPT-3 on CoreWeave GPUs. “The primary demand for GPUs is coming from a small number of very, very well-resourced organizations that put a very high premium on speed and on having the latest and greatest thing.”

And that’s leading them to make these scratch-my-back deals that are literally set up to help them help one another. Thanks in large part to the big deals, valuations are skyrocketing, and the billionaire founders, executives and investors tied to the massive AI data center buildout have benefited the most. By Forbes’ count, 20 of the most notable billionaires tied to the explosive growth in AI infrastructure spending have already added more than $450 billion to their fortunes since January 1.

Oracle cofounder and chief technology officer Larry Ellison is the biggest winner, up $140 billion in the past year thanks to a 73% share jump (vs. the S&P’s 15% increase), due in part to projections that revenue from cloud infrastructure, largely to power AI, would soar from $18 billion this year to $144 billion over the next four years. Nvidia cofounder and CEO Jensen Huang’s fortune has increased by $47 billion this year as shares of his chipmaking giant have risen by 40%, while Michael Dell has gotten $35 billion richer thanks to his stake in Dell (up 39%) and his estimated stake in Broadcom (up 49%), with both companies providing many of the parts that go into AI data centers.

The biggest gainers by percentage of net worth are cloud computing firm CoreWeave’s five billionaires. CoreWeave’s shares are up 250% since its March IPO, nearly tripling the net worths of its four billionaire cofounders (Michael Intrator, Brian Venturo, Brannin McBee and Peter Salanki) and early investor Jack Cogen. To finance its infrastructure buildouts as quickly as possible, CoreWeave has raised some $29 billion in debt—but says that materially all of it is priced into multiyear contracts (average length: four years) with the likes of Meta, Microsoft and OpenAI.

SoftBank’s Masayoshi Son and Russian search giant Yandex’s founder Arkady Volozh are also big beneficiaries. Their fortunes have increased by 142% and 166% this year, respectively. In April, SoftBank announced $40 billion in additional funding for OpenAI—though $30 billion of that is contingent upon OpenAI’s conversion to a for-profit by December. Volozh’s publicly known wealth now sits in Nebius, which builds AI data centers and rents GPUs to the likes of Microsoft. (The tech giant signed a $17 billion deal with Nebius last month.) Nebius is up 340% this year, pushing Yandex’s former chief information security officer Vladimir Ivanov into the billionaire ranks alongside his former boss for the first time thanks to his $1.2 billion Nebius stake.

Got a tip? Contact Phoebe Liu at pliu@forbes.com or phoebe.789 on Signal.

As valuations climb, the companies and investors project confidence that they’re not susceptible to the substantial risk this all blows up. Big tech firms like Oracle, Microsoft and Google that are spending big on AI infrastructure have cash cow businesses that help finance it all. Still, Oracle is carrying the most debt it’s ever had—it issued another $18 billion in debt in September, and the S&P’s credit rating bureau downgraded its outlook for the company to “negative” in July, citing concerns about free cash flow. “Growth prospects are strong, but risks include potential data center overcapacity should AI compute demand slow over time, customer and supplier concentration risk, and [a] still evolving competitive landscape,” S&P Global Ratings wrote in its September credit rating update. Big AI infrastructure lenders like Blackstone, which led a $7.5 billion debt financing for cloud computing firm CoreWeave, claim they’ve made the contracts airtight so customers can’t easily walk away. Adds CoreWeave CEO Michael Intrator: “I’m selling it to Microsoft. Microsoft is going to pay its bills.” (But OpenAI, which signed $22 billion in compute contracts with CoreWeave this year, making it CoreWeave’s largest customer behind Microsoft and Meta, has a lot less cash.)

OpenAI and other private companies are also being rewarded with big checks and lofty valuations. Investors valued OpenAI at $157 billion in October 2024, $300 billion in August and now $500 billion, the highest ever valuation for a private company. The company’s cofounder and CEO Sam Altman has long insisted that he doesn’t have a material stake in the company (which is currently a non-profit), though he’s a billionaire thanks to his other investments—including an $800 million stake in pre-revenue nuclear power company Oklo, which could power AI data centers if its technology ever gets up and running.

OpenAI rival Anthropic, which announced enterprise deals with IBM and Deloitte this week and has deep partnerships with (and large investments from) Amazon and Google, was valued at $183 billion in a September fundraise, up from some $18 billion in late 2024. Anthropic’s seven cofounders now hold stakes worth $3.7 billion apiece, after not being billionaires a year ago. In hopes that one of these companies will create the machine-god-esque artificial general intelligence, investors have also been pouring money into pre-revenue AI labs at astounding valuations, including former OpenAI chief technology officer Mira Murati’s Thinking Machines, which raised $2 billion at a $12 billion valuation in July, and former OpenAI chief scientist Ilya Sutskever’s Safe Superintelligence, which raised $2 billion at a $32 billion valuation in April.

“The problematic thing all comes down to one issue, which is, who’s gonna take residual risk on the technology?” says Chris Moon, managing director at DigitalBridge, which manages $106 billion in assets related to digital infrastructure, including AI data centers.

Whether all this growth is sustainable hinges on companies’ ability to turn AI innovation into sustainably profitable businesses—not least OpenAI, which doesn’t yet know how it’ll make a profit or raise all that money. In any case, some have started to cash out. OpenAI reportedly completed a $6.6 billion employee share sale last week, while CoreWeave’s billionaire insiders have already sold more than $1.3 billion of shares combined and Nvidia’s Huang sells stock amost daily.

Additional reporting from Rashi Shrivastava and Emily Garcia.