Picks and Shovels: How the AI Boom Made the Suppliers Richer Than the AI Companies
The Million Dollar Question: Michael Dell added roughly $107 billion to his fortune in 2026 — a gain Billionaires.Africa reported was “second only to” one other person’s. Whose?
A) Jensen Huang B) Elon Musk C) Larry Ellison D) Mark ZuckerbergRead on for the answer.
Dell Technologies turns 42 years old this year. It sells computers. In 2026 it made its founder one of the biggest wealth gainers alive — bigger than the chip designer who makes the hardware behind every large AI model, bigger than the cloud giant whose stock briefly made its CEO the richest person on Earth, and by a wide margin bigger than the man running the company that actually built ChatGPT. That last part isn’t a rounding error. It’s the whole story.
What it is
During the California Gold Rush, most of the people who went looking for gold didn’t find much of it. The reliable money went to the merchants who sold the miners what they needed to keep looking: pickaxes, shovels, tents, and — in the case of a Bavarian immigrant named Levi Strauss — a sturdy pair of riveted denim pants. Samuel Brannan, often cited as California’s first millionaire, didn’t pan for gold either; he bought up every shovel and pan in the region and resold them at a markup once news of the discovery spread. “Picks and shovels” became shorthand for a simple investing idea: when an industry is in a frenzy, don’t bet on who strikes gold. Bet on who’s selling the tools.
The AI boom has run the same play at a scale no gold rush ever approached. The “miners,” in this analogy, are the AI labs — OpenAI, Anthropic, xAI, and a handful of others — racing to build ever-larger models, spending tens of billions of dollars a year on computing power they mostly rent or lease. The “shovel sellers” are the companies that design the chips, build the servers, run the data centers, and keep the lights and cooling on. Chip design (Nvidia), server assembly (Dell, among others), and cloud capacity (Oracle, CoreWeave) sit in the second category. So, this year, do several of the single largest personal fortune gains on the planet — not at the AI labs, but one layer down, in the infrastructure underneath them.
Who’s getting rich
Start with the clearest case. Michael Dell co-founded his company from a University of Texas dorm room in 1984 selling upgraded PCs by mail order. In 2026, Dell Technologies’ AI-optimized server revenue rocketed 757% to $16.1 billion in a single fiscal quarter, and the stock roughly tripled from where it started the year, according to Fortune. Because Dell personally owns roughly 40% of the company he founded, that surge flowed straight into his own net worth. He started 2026 at an estimated $141 billion, ranked 13th in the world, according to American Bazaar; by September 4 he’d added roughly $107 billion for the year, a gain Billionaires.Africa reported was “second only to Elon Musk’s.” Five days later, on September 9, his fortune hit $267.7 billion and he passed Jeff Bezos to become the world’s third-richest person, per American Bazaar. By mid-September, Fortune put the figure at $262 billion — “nearly doubled” year over year — and Forbes’ real-time tracker had him at $266.1 billion and world #4 as of October 1.
Jensen Huang’s Nvidia designs the chips that make most of this possible, and he’s been wealthy far longer than this particular boom — but even he was still climbing in 2026, worth an estimated $202.5 billion and ranked #7 on Forbes’ real-time list as of October 4, with Nvidia itself becoming one of the first companies ever to cross a $5 trillion market capitalization, per Yahoo Finance.
Larry Ellison got there first and biggest, if only for a day. On September 10, 2025, Oracle announced quarterly results built around what NBC News reported was a cloud-services backlog exceeding $450 billion, driven in large part by multibillion-dollar computing contracts with Nvidia and OpenAI. Oracle’s stock jumped as much as 43% that day. Ellison’s net worth, according to Storyboard18, hit roughly $393 billion, edging past Elon Musk’s $385 billion to make the 81-year-old Oracle co-founder — for a few hours — the richest person alive. He didn’t build a chatbot. He sold the company that rents computers to the people who did.
Then there’s the newest tier: companies that don’t design chips or write software so much as buy, lease, and resell raw computing capacity to the AI labs that need more of it than anyone can build fast enough. CoreWeave, originally a cryptocurrency-mining operation, pivoted to renting out Nvidia GPU clusters and went public in 2025. Its CEO, Michael Intrator — a former hedge fund manager — saw his net worth double from $5 billion to $10 billion in twelve days in June 2025, according to Fortune, which pointedly noted that neither he nor his co-founders — Brian Venturo (then $6.4 billion) and Brannin McBee (then $4.7 billion) — came up through the tech industry at all.
Why the suppliers, and not the labs
Here’s the part that should feel strange, and is worth sitting with: the people actually building the AI models that triggered this entire boom are, with a handful of exceptions, not showing up on the list above.
The clearest case is Sam Altman, OpenAI’s CEO. As of this writing, Altman owns no equity in OpenAI, a fact confirmed by Fortune, which reported that OpenAI’s board had discussed — but not committed to — giving him a stake, and that he had told staff there were “no current plans” to do so. Altman is a billionaire, but his money comes almost entirely from roughly 80 angel investments made before and alongside his OpenAI tenure, not from the company whose product reshaped the entire technology industry. Anthropic’s founders hold equity in the company they built, but Anthropic remains private; whatever their paper stakes are worth doesn’t show up on a real-time billionaire tracker the way a NYSE-listed stock does, because there’s no public market minting a fresh price on it every second.
That’s the mechanical reason the money landed where it did. Dell Technologies, Nvidia, Oracle, and CoreWeave are all publicly traded. Their revenue from selling hardware and cloud capacity is real, recognized, and reported every quarter — and when it beats expectations, the stock repricing is immediate and the founder’s or major shareholder’s net worth is repriced right along with it. OpenAI, Anthropic, and xAI, by contrast, are privately held and by most public reporting still burning enormous amounts of cash on the very compute that Dell, Nvidia, and Oracle are selling them. The AI labs are the customers in this transaction. The infrastructure companies are the ones sending the invoice — and collecting on it now, rather than someday.
How the fortunes actually move
It’s worth understanding the mechanics, because they explain both the scale of these gains and how fast they can reverse. A fortune like Dell’s isn’t cash sitting in an account; it’s a percentage ownership stake in a public company, multiplied by whatever the stock is trading at the moment someone checks. Dell owns about 40% of Dell Technologies. When the stock roughly tripled over 2026 on AI-server demand, his paper net worth didn’t just go up — it went up by close to 40% of that entire increase in the company’s total market value, because that’s his slice.
This is also exactly how Forbes’ and Bloomberg’s real-time billionaire trackers work: take the public share count and ownership stake, multiply by the live stock price, publish the number, and update it continuously during trading hours. It’s a legitimate and widely used methodology — but it means the number is only ever a snapshot, built entirely on a stock price that can move 10% or 40% in a single session on one earnings report, the way Oracle’s did in September 2025. Compare that to a privately held AI lab, where the “valuation” is whatever the most recent funding round priced it at, sometimes months or a year stale, with no daily market check at all. Both numbers are estimates. Only one of them updates every few seconds.
What it costs
The money flowing to the suppliers has to come from somewhere, and where it comes from is the AI labs’ own capital spending — along with that of the hyperscalers (Microsoft, Google, Amazon, Meta) building out AI infrastructure for their own products. Oracle’s $450 billion-plus cloud backlog, cited by NBC News, is essentially a multi-year promise from customers — including OpenAI — to keep paying for computing capacity at a scale few companies have ever contracted for. Broadcom’s chief executive has told investors he expects the company’s AI chip sales alone to top $100 billion by 2027, according to Bloomberg reporting on the company’s guidance. And combined 2026 AI capital-spending commitments from just four hyperscalers — Amazon, Microsoft, Alphabet, and Meta — had already topped $700 billion, according to Yahoo Finance, which cited Meta alone raising its full-year capex guidance to $125 billion–$145 billion.
Industry-wide capital spending on AI chips, servers, data centers, and the power to run them is now measured in the hundreds of billions of dollars a year, funded by a mix of AI-lab revenue, hyperscaler cash flow, and — increasingly — debt. Every dollar of that capex is a dollar of potential revenue for somebody selling picks and shovels: a chip designer, a server assembler, a capacity landlord. It’s also a dollar of risk sitting on the balance sheet of whoever borrowed to spend it, a tension the market has not yet had to fully resolve. If AI-lab revenue ever stops justifying this pace of spending, the hundreds of billions of dollars currently flowing to suppliers each year is the first number that shrinks — which is exactly why the fortunes built on it move as fast on the way down as they did on the way up.
The scale of all this shows up at the top of the wealth pyramid as a whole. The Altrata Billionaire Census 2026, released August 27, 2026, counted 3,795 billionaires worldwide — up 8.2% from the year before, the fastest annual growth in five years — with combined wealth of $15.1 trillion, a 12.8% increase. Altrata’s researchers singled out AI investment as a major driver, finding that companies making “meaningful AI investments” saw valuations climb roughly 23% faster than companies that didn’t, over 2024 and 2025 combined.
Hidden costs and tradeoffs
The flip side of a fortune that can gain $107 billion in eight months is that it can give a meaningful chunk of that back just as fast — and Dell’s own numbers show it happening in real time. His estimated net worth went from $267.7 billion on September 9 to $262 billion by September 14, before Forbes’ real-time tracker had him back up at $266.1 billion on October 1. None of those moves reflect anything Michael Dell did personally in that window; they reflect Dell Technologies’ stock price ticking around on a given trading day. (For more on why a headline net-worth figure is a snapshot rather than a bank balance, see Net Worth Is Not Net Worth.)
There’s a sharper cautionary tale one tier down the hardware-supplier chain. In 2024, server maker Super Micro Computer — a smaller player in the same AI-infrastructure business as Dell — saw its founder’s paper wealth collapse by roughly two-thirds in a matter of months after an accounting scandal delayed its financial filings and triggered a Nasdaq delisting threat. The lesson generalizes: a fortune built on a hardware-cycle stock price is only as durable as the next quarterly report, the next audit, and the next sign that AI capital spending is still accelerating rather than plateauing. Even Michael Dell, selling an unambiguously real and currently booming product, is pragmatic enough to be quietly offloading some of his own real estate at a loss — he’s auctioning a never-lived-in Boston penthouse bought for $10.9 million in 2016 with an opening bid of just $6.5 million, an auction that closes October 29, 2026, according to Billionaires.Africa and corroborated by Bisnow’s coverage of the listing. Even at $100M–$300M+ net worth, not every asset gets held for the win.
What people get wrong
The most common mistake is assuming “AI billionaire” automatically means someone who founded or runs an AI lab. In the current boom, most of the largest individual wealth gains have gone to people running decades-old hardware and infrastructure companies — Dell Technologies, Nvidia, Oracle — that happened to be positioned exactly where the spending landed, plus a small number of infrastructure-rental upstarts like CoreWeave. The AI labs themselves, meanwhile, often aren’t profitable, aren’t public, and in at least one marquee case — Sam Altman at OpenAI — the person running the company doesn’t hold equity in it at all.
The second mistake is treating these as brand-new fortunes created from nothing, the way a startup IPO mints new wealth. Dell and Oracle are 42- and 48-year-old companies respectively; Nvidia has been public since 1999. What’s new isn’t the company or the founder — it’s the growth story Wall Street is pricing in, and how fast that repricing happened. CoreWeave is the real exception here: a genuinely new company, repurposed from crypto mining, whose founders went from relatively obscure to multibillionaires inside a couple of years with essentially no prior technology-industry track record.
The third mistake is reading these numbers as settled fact rather than live estimates. Every figure in this piece — Dell’s $266.1 billion, Huang’s $202.5 billion, Ellison’s brief $393 billion — is a real-time mark-to-market calculation that was accurate at the moment it was reported and has almost certainly changed by the time you’re reading this.
Bottom line
So: who gained more than Michael Dell’s roughly $107 billion in 2026? Elon Musk — the answer is B. It’s a useful fact mostly because of what it reveals about who else didn’t make that list. Sam Altman, who runs the company most people would point to first if asked “who’s getting rich from AI,” isn’t on it, because he doesn’t hold equity in OpenAI. Dario Amodei at Anthropic isn’t on it either, because Anthropic is private and its stock doesn’t reprice every morning. The people who are on it — Dell, Huang, Ellison, and a trio of former finance professionals running a GPU-leasing company — sell the chips, boxes, and computing capacity that the AI labs are paying for, often at a loss, to build the thing everyone’s actually talking about. Picks and shovels, it turns out, is still the best bet in the room. Whether that holds once the AI labs themselves turn a profit — or once the capital-spending cycle cools — is the open question the next few years of this list will answer.
Related reading: SpaceX, OpenAI, Anthropic, and the Next Gold Rush of Tech Wealth · Tech Wealth: How Founders and Investors Live Differently · Equity Compensation: RSUs, ISOs, and the Tech Wealth Engine · Net Worth Is Not Net Worth: How Billionaire Fortunes Evaporate on Paper · Anatomy of the Forbes 400: Who’s Actually On the List
