The top money managers of Wall Street billionaires have picked their favorite AI stocks for the new year — and Nvidia didn’t make the cut.
Over the past two years, no trend has put more pep in Wall Street’s step than the rise of artificial intelligence (AI). The ability of AI-driven software and systems to become more proficient at their tasks, and to learn new jobs without the need for human intervention, gives this technology a virtually unlimited long-term ceiling.
While growth estimates vary widely, PwC analysts see an addressable AI market of $15.7 trillion by 2030. According to PwC’s Determine the priceincreased productivity will increase global gross domestic product by $6.6 trillion, while the side effects of consumption will add another $9.1 trillion.
Image source: Getty Images.
This predicted outperformance and the sky-high ceiling for artificial intelligence are not lost on Wall Street and its top investors. Quarterly filed Form 13Fs allow investors to track which AI stocks top money managers have been buying and selling.
Based on the latest round of 13Fs, which cover trading activity through the end of September, there are three AI stocks that billionaire asset managers will clearly want to own heading into 2025.
Broadcom
The First AI Stock Billionaires Can’t Get Enough As We Enter a New Year, Is a Network Solutions Specialist Broadcom (AVGO 0.15%). Based on the 13Fs for the quarter ended September, billionaires Philippe Laffont of Coatue Management (1,488,666 shares purchased) and Stanley Druckenmiller of Duquesne Family Office (239,980 shares purchased) were buyers.
Just like Nvidia (NVDA 3.59%) While it has become the undisputed top option as a provider of graphics processing units (GPUs) for companies looking to build out AI-accelerated data centers, Broadcom has become a major provider of networking solutions within those data centers. The company’s Jericho3-AI fabric can connect up to 32,000 GPUs, which is essential for maximizing GPU computing capabilities and reducing tail latency.
Furthermore, Broadcom is ideally positioned to capitalize on enterprise demand for its custom AI chips. CEO Hock Tan thinks the company’s AI revenues could rise to between $60 billion and $90 billion by fiscal 2027, up from the $12.2 billion reported in fiscal 2024 (the fiscal year ended Nov. 3 ). Demand from the company’s key hyperscale customers should drive this growth.
Perhaps the most attractive aspect of Broadcom to Laffont and Druckenmiller is that it is much more than just an AI stock. It is a top supplier of wireless chips and accessories used in smartphones, supplies a laundry list of optical sensors to the industrial sector and has a range of cybersecurity solutions. If an AI bubble were to form, Broadcom would be far better suited than Nvidia to weather the storm.
The big question for these two billionaires is, “Can Broadcom maintain its trillion-dollar valuation?” While sustainable double-digit growth seems likely, Broadcom trades at a price-to-sales multiple Good above its historical average. We may see the company’s stock flounder until its valuation becomes more bearable.
Image source: Getty Images.
Taiwanese semiconductor manufacturing
A second AI stock that billionaire asset managers can’t stop buying before 2025 is the world’s largest chipmaker Taiwanese semiconductor manufacturing (TSM 2.67%). During the third quarter, billionaire Chase Coleman of Tiger Global Management picked up 564,090 shares, while Duquesne’s chief Stanley Druckenmiller picked up 57,355 shares.
Taiwan Semi is being counted on by leading companies in the AI industry, including Nvidia, to dramatically increase GPU production. Based on recently updated targets following Donald Trump’s victory in November, Taiwan Semi aims to achieve monthly chip-on-wafer-on-substrate (CoWoS) capacity of 35,000 units by 2024, 75,000 units by 2025 and 135,000 units by 2026. CoWoS is a necessity for packaging the high-bandwidth memory that supports AI-accelerated data centers.
Taiwan’s semiconductor manufacturing should also continue to benefit from long-term lags in AI chips. As long as the AI GPU shortage persists, the company’s operating cash flow can remain highly predictable.
The big question billionaires Coleman and Druckenmiller will have to ask themselves is: “How much will Trump’s trade policies affect the company?” The incoming president’s U.S. focus and expected dependence on tariffs could create challenges for Taiwan Semi, which has between 80% and 90% of its production capacity in Taiwan. Export restrictions on AI chips and equipment to China, implemented by the Biden administration, could also pose a challenge.
Like Broadcom, Taiwan Semiconductor Manufacturing is no longer the phenomenal fundamental bargain it has been for years. The price-earnings ratio (P/S) is 45% above the average price-earnings ratio over the last five years, while the price-earnings ratio (P/E) of 23 represents the highest value since 2020. It’s a somewhat aggressive valuation for a very cyclical industry.
Amazon
The third artificial intelligence stocks selected by billionaires who can’t stop buying before the new year are e-commerce giants Amazon (AMZN 1.37%). Four billionaires were buyers in the quarter ended September, including (total shares purchased in brackets):
- Stephen Mandel of Lone Pine Capital (1,033,987 shares)
- Philippe Laffont of Coatue Management (496,218 shares)
- Larry Robbins of Glenview Capital Management (125,000 shares)
- Chase Coleman of Tiger Global Management (94,408 shares)
Amazon’s AI bands are primarily based on usage. Amazon Web Services (AWS) is the world’s leading cloud services infrastructure platform, aggressively integrating generative AI solutions. Generative AI on AWS can help companies build AI applications, deploy virtual chatbots and AI assistants, and build/run large language models.
Of Amazon’s numerous operating segments, none is more important for generating cash flow or profit than AWS. Through the first nine months of 2024, AWS accounted for 17.5% of Amazon’s net revenue, but nearly 62% of its operating income. The juicy margins that typically come with cloud subscriptions will play a key role in increasing the company’s cash flow over time.
Amazon is also developing its own AI chips, known as the Trainium2 and Inferentia. While Amazon already uses Nvidia’s top GPUs, and its own chips are unlikely to match Nvidia in terms of computing speed, the Trainium2 and Inferentia should be significantly cheaper and easier to access than Nvidia’s sought-after hardware.
As with Broadcom and Taiwan Semi, the big concern for Mandel, Laffont, Robbins and Coleman is whether Amazon stock is still a buy after hitting an all-time high. While traditional fundamental tools such as the price-to-earnings ratio suggest that Amazon is more than fully valued, the company’s price-to-cash flow ratio implies that the company still offers upside potential. At 16 times estimated 2025 cash flow, Amazon is still well below the multiple of 23 to 37 times cash flow that investors regularly paid for its stock in the 2010s.
