If You Missed Palantir or Nvidia. This is Far Bigger.
Summary
Tom identifies the next phase of the AI cycle as agentic AI and robotics, moving beyond simple chatbots to autonomous systems that perform tasks independently. Tom explains that these systems require significantly more memory and context, which shifts hardware requirements toward a much higher ratio of CPUs per GPU. Tom dismisses claims that AI is a bubble, citing data that major data centers are already profitable, generating billions in annual returns on investment.
Tom highlights the Jevons Paradox, suggesting that as AI technology becomes more efficient and cheaper, global demand will skyrocket rather than diminish. Tom believes that while the short-term market may be stretched, the long-term potential for infrastructure providers is massive. Tom advises using a double-down dollar-cost averaging strategy to build positions in high-quality companies over the next year.
Mentioned Stocks
Reasoning: Tom notes that while Nvidia is a top-tier company, it has already realized a 15x return since his previous call. Tom suggests looking for the next player in the second wave of AI infrastructure rather than chasing previous gains.
Reasoning: Tom highlights that AMD is trading at its cheapest price-to-sales ratio (17x) since 2022 despite explosive growth in operating income and cash reserves. Tom argues that the shift to agentic AI will require a 1:1 or 1:2 CPU-to-GPU ratio, drastically increasing AMD's total addressable market. Tom provides five-year price predictions of 158% in a bear case, 534% in a medium case, and 1,500% in a bull case scenario.
Reasoning: Tom mentions Intel as a key CPU player currently down 40% from its June highs. Tom believes you can't go wrong with it, but expresses a preference for AMD's fundamentals.