For investors, the quarterly release of Form 13Fs with the Securities and Exchange Commission (SEC) is like Christmas all over again.
A 13F provides investors with a detailed snapshot of what stocks Wall Street's top money managers bought and sold in the latest quarter (in this instance, the fourth quarter). In other words, it offers an inside look at what stocks, industries, and trends are piquing the interest of the brightest asset managers.
Image source: Getty Images.
There's perhaps no bigger intrigue at the moment than knowing what Wall Street's top money managers are doing with artificial intelligence (AI) stocks. AI, which involves using software and systems to handle tasks that would typically be overseen by humans, could add an estimated $15.7 trillion to global gross domestic product by 2030, according to a report released last year by PwC.
Based on the latest round of 13F filings, billionaire investors were eager to pare down their stakes in two ultra-popular AI stocks and simply couldn't stop buying shares of another brand-name AI-inspired company.
Widely owned AI stock No. 1 that billionaires are decisively selling: Nvidia
The first artificial intelligence stock that prominent billionaire investors have been kicking to the curb is none other than the infrastructure backbone of the AI movement, Nvidia (NVDA 6.18%). During the December-ended quarter, eight billionaires reduced their fund's respective stakes in this semiconductor giant, including (total shares sold in parenthesis):
- Israel Englander of Millennium Management (1,689,322 shares).
- Jeff Yass of Susquehanna International (1,170,611 shares).
- Steven Cohen of Point72 Asset Management (1,088,821 shares).
- David Tepper of Appaloosa Management (235,000 shares).
- Philippe Laffont of Coatue Management (218,839 shares).
- Chase Coleman of Tiger Global Management (142,900 shares).
- John Overdeck and David Siegel of Two Sigma Investments (30,663 shares).
If you're wondering why these top investors would lighten their load on Wall Street's hottest megacap stock, increasing competition and margin cannibalization may be the answer.
For the moment, Nvidia's A100 and H100 graphics processing units (GPUs) dominate AI-accelerated data centers. Analysts at Citigroup have estimated that Nvidia's chips could account for a 90% share of GPUs deployed in high-compute data centers this year. But competition is picking up. Advanced Micro Devices will be increasing the rollout of its MI300X GPU this year, while Intel is debuting its Gaudi3 generative AI software chip as a direct competitor to the H100 later this year.
Internal competition is a potential problem, too. Microsoft (MSFT 1.82%) and Meta Platforms are Nvidia's respective No. 1 and No. 2 in total spending. But even though these companies are spending big bucks with Nvidia, they're hard at work developing their own AI chips. Sooner than later, Microsoft and Meta's reliance on Nvidia should lessen.
Nvidia's expansion could also, ironically, be its undoing -- at least when it comes to the company's gross margin. The bulk of Nvidia's data-center growth in fiscal 2024 (ended in late January) was the result of A100 and H100 GPU scarcity. When a product is scarce and in high demand, the seller typically commands exceptional pricing power. As Nvidia ramps up production of its high-powered GPUs, it could cannibalize its own stellar pricing power.
Widely owned AI stock No. 2 that billionaires are decisively selling: Microsoft
The other AI stock that billionaires have thrown out with the bath water is the world's largest publicly traded company by market cap, Microsoft. Despite investing aggressively in OpenAI (the company behind popular chatbot ChatGPT) and developing its own AI chips, seven genius billionaire investors were sellers during Q4, including (total shares sold in parenthesis):
- Ole Andreas Halvorsen of Viking Global Investors (3,024,399 shares).
- Steven Cohen of Point72 Asset Management (1,569,462 shares).
- Jim Simons of Renaissance Technologies (1,155,782 shares).
- Ken Griffin of Citadel Advisors (796,892 shares).
- Chase Coleman of Tiger Global Management (787,113 shares).
- Terry Smith of Fundsmith (705,498 shares).
- Dan Loeb of Third Point (210,000 shares)
Viking Global and Renaissance Technologies completely exited their fund's stakes in the world's largest public company. The reasoning behind these sales may boil down to a combination of history and valuation.
Over the past 30 years, there has been no shortage of next-big-thing investment trends. While some of these trends have made long-term investors notably richer (e.g., the advent of the internet and cloud computing), every next-big-thing investment has navigated its way through an initial bubble. This is to say that investors regularly overestimate the early uptake of new innovations or technologies, and AI is unlikely to be an exception.
Microsoft's valuation is another potential sticking point. Investors right now would be paying a multiple of 31 times forward-year earnings, which is a far cry from the year-end multiple of 16 to 25 times forward-year earnings investors were paying between 2014 and 2018.
The one thing that could ultimately make these billionaires regret their decision to sell is Microsoft's incredible cash flow. The perfect blend of high-margin cash flow from its legacy segments (e.g., Windows and Office), coupled with high-growth initiatives like cloud service infrastructure-platform Azure, afford Microsoft the luxury to take chances on the acquisition and innovation front.
Image source: Getty Images.
The artificial intelligence stock billionaires can't stop buying: Amazon
However, not all artificial intelligence stocks were off-limits by billionaire money managers during the December-ended quarter. E-commerce company Amazon (AMZN 3.43%), which is utilizing generative AI in a variety of ways to enhance voice interactions with Alexa, as well as help merchants create more compelling and tailored advertisements, was a popular buy for eight billionaires, including (total shares purchased in parenthesis):
- Ken Griffin of Citadel Advisors (4,321,477 shares).
- Jim Simons of Renaissance Technologies (4,296,466 shares).
- Chase Coleman of Tiger Global Management (947,440 shares).
- Ken Fisher of Fisher Asset Management (888,369 shares).
- John Overdeck and David Siegel of Two Sigma Investments (726,854 shares).
- Steven Cohen of Point72 Asset Management (462,179 shares).
- Israel Englander of Millennium Management (85,532 shares).
Although AI contributes to the lure of Amazon's various products and services, billionaires are most likely piling in because of the company's three rapidly growing ancillary segments.
Most consumers are probably familiar with Amazon because it has the most-dominant online marketplace in the United States. While this segment is responsible for a lot of revenue, online retail sales produce low margins. The lion's share of Amazon's cash flow can be traced back to its cloud-infrastructure service segment, Amazon Web Services (AWS), as well as subscription services and advertising services.
AWS is Amazon's superstar. It's the leading provider of cloud-infrastructure services globally and is approaching $97 billion in annual run rate sales. Not only is enterprise spending on cloud-infrastructure services still in its early innings, but the margins associated with cloud services put e-commerce margins to shame. It's not uncommon for AWS to account for more than half of Amazon's operating income.
The other catalyst that encouraged billionaires to mash the buy button for Amazon stock during Q4 might be its valuation. While the traditional price-to-earnings ratio is liable to have value investors running in the opposite direction, Amazon is historically cheap relative to its cash flow.
Cash flow tends to be a smarter way of valuing Amazon given its penchant for reinvesting most of its operating cash flow back into its logistics operations and various high-growth initiatives. After spending the entirety of the 2010s valued between 23- and 37-times year-end cash flow, investors can buy shares today for about 12 times consensus cash-flow estimates in 2025.
