While billionaire investor Lee Ainslie of Maverick Capital may not be a household name, he is a well-regarded investor who comes from the "Tiger Cub" lineage of portfolio managers who worked under legendary investor Julian Robertson's Tiger Management before branching out on his own. In the past, Ainslie has said he places great emphasis on strong management teams when making investments.
In the first quarter, he bought shares in a trio of hardware companies poised to benefit from the rise of agentic AI. With the proliferation of artificial intelligence (AI) agents, data centers are expected to see a big shift in the ratio of graphics processing units (GPUs) to central processing units (CPUs). In large language model (LLM) training, the ratio of GPUs to CPUs was 8:1, but with agentic AI, the ratio is expected to shift to 1:1, as CPUs are needed to provide sequential reasoning and the ability to work with tools required by AI agents.
Let's take a close look at three AI stocks Ainslie was buying and whether they still look like good investments.
Image source: Getty Images.
1. Nvidia
Nvidia (NVDA 5.93%) is Maverick's largest holding, representing more than 6% of the fund's portfolio, and Ainslie was adding to his position in Q1. While Nvidia has established itself as the undisputed king of AI model training with its GPUs, it also sees a big opportunity stemming from the rise of agentic AI and the need for more data center CPUs.
NASDAQ: NVDA
Key Data Points
The company has created its own ARM-based Vera CPUs designed specifically for AI agents. It recently called this a $200 billion opportunity and is selling stand-alone CPUs for the first time. It's also developed a complete rack system for agentic AI, which includes its GPUs, CPUs, language processing units (LPUs), and networking components in a fully integrated system.
As the market moves toward inference and agentic AI, Nvidia remains very well positioned.
2. AMD
Maverick took a new position in Advanced Micro Devices (AMD 11.01%) in Q1, and the company is one of the best-positioned in the hardware space for the rise of agentic AI. It is a leader in the data center CPU space, having consistently taken share away from rival Intel (INTC 11.92%), and AMD is seeing robust demand for its chips.
NASDAQ: AMD
Key Data Points
AMD is also just starting to roll out its newest high-performance CPUs designed for agentic AI, built with Taiwan Semiconductor Manufacturing's 2-nanometer process technology. The new chips, codenamed Venice, feature a whopping 256 core that act as isolated work zones and offer substantial memory bandwidth. This should help the company continue to take market share and realize higher pricing.
On top of that, AMD is also seeing solid momentum in the inference market with its GPUs, having gained large commitments from both OpenAI and Meta Platforms. Between its CPU opportunity and these deals, AMD is poised for tremendous growth moving forward.
3. Intel
NASDAQ: INTC
Key Data Points
Rounding out his hardware-agentic AI investments, Ainslie began a position in Intel. Intel has long been the CPU leader by volume, and the sudden surge in CPU demand helped propel 22% revenue growth in its data center and AI (DCAI) segment in Q1. With data center CPU demand skyrocketing, the company should benefit from this trend. It could also have an advantage by owning its own manufacturing facilities, given that CPU supply is expected to remain tight.
That said, Intel is not high on my list of AI stocks to own. Its overall revenue growth is still modest, up only 7% in Q1, and the stock's more than fivefold increase over the past year has taken it from a very cheap stock to an expensive one. Meanwhile, the company still has issues with its foundry business, and its Diamond Rapids server CPUs are reportedly delayed until next year, highlighting Intel's continued struggles with innovation.
