For the last two and a half years, no trend has captivated more attention or capital from investors than artificial intelligence (AI). With AI, software and systems are capable of making split-second decisions without the need for human intervention.
This is a technology with seemingly limitless long-term potential. With AI offering some form of utility in virtually all industries around the globe, the analysts at PwC pegged its worldwide impact on gross domestic product by 2030 at $15.7 trillion. A figure this massive ensures there will be a long list of winners.
No company has been a more direct beneficiary of the evolution of AI than Nvidia (NVDA 1.45%), whose market cap skyrocketed by more than $3 trillion in less than two years.
But what's interesting is how Wall Street's most prominent billionaire money managers have approached Nvidia stock. Specifically, billionaire investors have been selling shares of Nvidia hand over fist... with one notable exception.
Image source: Getty Images.
Billionaire Chase Coleman continues to load up on Nvidia
No later than 45 calendar days following the end to a quarter, institutional investors overseeing at least $100 million in assets are required to file Form 13F with the Securities and Exchange Commission. This filing details for investors which stocks Wall Street's smartest money managers were buying and selling in the latest quarter.
Whereas most billionaires have been persistent sellers of Nvidia stock (a topic I'll touch on a bit later), Tiger Global Management's billionaire chief Chase Coleman has been the exception. Accounting for Nvidia's historic 10-for-1 forward split that was enacted in June 2024, Coleman's fund has expanded its stake in the company from 9,683,550 shares, at the end of 2023, to 10,967,550 shares, as of the end of March 2025.
Coleman's investment strategy has historically focused on a mix of large- and small-cap growth stocks, with a tendency to lean into next-big-thing trends and innovations. The rise of artificial intelligence certainly fits the mold of a game-changing trend that would interest Tiger Global's billionaire investor.
Chase Coleman choosing to expand his fund's exposure to Nvidia likely has to do with one or more of the following four factors.
First, Nvidia's share of the graphics processing unit (GPU) market in high-compute data centers can be described as monopoly like. Businesses have lined up to purchase Nvidia's Hopper (H100) and successor Blackwell GPUs, and there's been no evidence that demand for these products has slowed. Wall Street tends to reward companies with perceived-to-be sustainable moats.
Secondly, AI-GPU scarcity has been working in Nvidia's favor. Even with Wall Street's AI darling and other direct competitors pushing out as many GPUs as possible, the chip fabrication supply chain is currently maxed out. With demand heavily outpacing the supply of AI-GPUs, Nvidia has been able to charge a triple-digit percentage premium for Hopper and Blackwell, relative to some of its direct rivals. This helped to lift Nvidia's gross margin above 70%.
The third factor that likely played a role in enticing Coleman to expand his fund's Nvidia stake is CEO Jensen Huang's ambitious innovation timeline. Huang is overseeing the debut of a new advanced AI chip on an annual basis. Blackwell Ultra (second-half of 2025), Vera Rubin (second-half of 2026), and Vera Rubin Ultra (second-half of 2027) are next in line after the ultra-popular Blackwell chip. Nvidia looks to be in no danger of ceding its clear-cut compute advantages in AI-accelerated data centers.
Fourth and finally, Nvidia stock tumbled at times during the first quarter, which may have brought its forward-year earnings multiple down to a level that intrigued Coleman.
Image source: Getty Images.
Most billionaire money managers have been persistent sellers of Nvidia stock -- here's why
However, Chase Coleman's actions are out of the ordinary among billionaire fund managers. Since Nvidia stock kicked off its historic run-up, most billionaires have pared down their exposure to the AI colossus or completely exited their respective positions:
- Stanley Druckenmiller of Duquesne Family Office sold all 9,500,750 shares of Nvidia stock between June 30, 2023 and Sept. 30, 2024.
- Stephen Mandel of Lone Pine Capital dumped all 6,416,490 shares between June 30, 2023 and June 30, 2024.
- David Tepper of Appaloosa shed 97% of his fund's stake (9.95 million shares sold) in Nvidia since Sept. 30, 2023.
- Philippe Laffont of Coatue Management has sent over 41.2 million shares to the chopping block since March 31, 2023 and reduced his fund's holdings by 83%.
- Israel Englander of Millennium Management, who regularly hedges with put and call options, has trimmed nearly 28.3 million shares (a 75% reduction) since Sept. 30, 2023.
This persistent selling activity may well be explained by nothing more than simple profit-taking. We've never witnessed a megacap stock gain $3 trillion in market cap in less than two years before, which has offered plenty of opportunity for billionaires to cash in their chips.
But there may be more to this consistent selling than just a desire to take profits.
For example, competition is inevitable in AI-accelerated data centers. Even if Nvidia continues to be the preferred option in high-compute data centers, it's unlikely to retain its monopoly like market share as direct competitors ramp up their GPU production.
Growing competition looks to be weighing on Nvidia's superior gross margin. NVDA Gross Profit Margin (Quarterly) data by YCharts.
What's potentially more worrisome is that most members of the "Magnificent Seven" are internally developing AI chips to use in their data centers. While these GPUs are no threat to Nvidia's compute advantages, they are notably cheaper and not backlogged due to demand. The mere presence of these chips means less future opportunity for Nvidia's hardware, as well as less in the way of premium pricing power.
There's also the very real possibility of Jensen Huang's accelerated innovation timeline hurting, not, helping, his company. Though bringing a new chip to market annually will cement its compute advantage, it has the potential to quickly depreciate the value of prior-generation chips. Wagering on Nvidia's largest customers to upgrade their hardware on an annual or near-annual basis is risky and may not pay off.
Perhaps the biggest concern for a majority of billionaire money managers selling Nvidia stock is the prospect of history rhyming. Every game-changing trend and innovation for more than three decades has endured an eventual bubble-bursting event early in its expansion. This is due to investors overestimating the mainstream adoption and/or utility of a new technology or innovation.
Considering that most businesses aren't generating a positive return on their AI investments, nor have they optimized their AI solutions, it looks to be a matter of time before the AI bubble bursts. If history were to rhyme, no company would be expected to take it on the chin more than Nvidia.
