Should Nvidia Worry About AMD and Google Breathing Down Its Neck in 2026?

AMD has laid out an aggressive growth target — >35% revenue CAGR over the next 3–5 years and >60% CAGR for its data‑center business — and is backing it with material customer wins: OpenAI will deploy AMD Instinct MI450 GPUs in H2 2026 (with warrants to buy up to 160 million shares), Oracle plans an AMD‑powered AI supercluster in Q3 2026, Cohere is integrating AMD GPUs, and the U.S. DOE awarded AMD gear for its Lux and Discovery supercomputers. Meanwhile Google Cloud’s TPUs are making inroads — Apple trained its Apple Intelligence models on TPUs, Anthropic is committing tens of billions to TPU capacity in 2026, Meta is in talks to use TPUs from 2027, and Google’s new Ironwood TPU is reportedly four times more powerful than its prior top chip. Collectively these moves heighten competitive pressure on Nvidia’s GPU dominance and imply the AI‑compute market can support multiple winners; Nvidia likely retains near‑term leadership, but execution on AMD contracts, OpenAI warrant exposure and accelerating TPU adoption are key potential catalysts for future share shifts.

Analysis

AMD has presented an aggressive growth target—greater than 35% revenue CAGR over the next three to five years and more than 60% CAGR for its data-center business—and the company has secured concrete customer commitments that underpin that ambition. Notable contract wins cited include OpenAI’s planned deployment of AMD Instinct MI450 GPUs in H2 2026 (with warrants to buy up to 160 million AMD shares), Oracle’s AI supercluster scheduled for Q3 2026, Cohere integrations, and U.S. Department of Energy selections for the Lux and Discovery supercomputers. Alphabet’s Google Cloud is increasing competitive pressure via TPUs, with Apple training its Apple Intelligence models on TPUs, Anthropic committing “tens of billions” to TPU capacity in 2026, Meta in talks to adopt TPUs from 2027, and the new Ironwood TPU claiming four times the previous performance. The author concludes Nvidia is likely to retain leadership in the near term—supported by Nvidia’s own messaging that it continues to supply Google—but the market can support multiple winners; key risks to monitor are execution and timing of AMD deployments, the materiality of OpenAI’s warrants, and accelerating TPU adoption by hyperscalers.

Key Decisions for Investors

• Consider a measured, tactical position in AMD to capture upside if it executes on the H2 2026 and 2026 customer rollouts, but size exposure conservatively given execution and timing risk
• Monitor the OpenAI warrants (up to 160 million shares) as a potential catalyst for AMD equity ownership changes or dilution and adjust positions if conversion terms or timing become material
• Add selective exposure to Alphabet as a strategic TPU play given Apple, Anthropic, and potential Meta adoption, while tracking commercial traction for Ironwood
• Maintain a core overweight or hold in Nvidia for near-term market leadership due to platform breadth, but set clear sell/hedge triggers tied to confirmed market-share losses or accelerated AMD/TPU deployments
• Use hedges or position limits rather than full-conviction bets and watch milestone-driven indicators (AMD deployments start dates, Oracle/Cohere/DOE rollouts, Anthropic/Meta TPU purchases) before increasing position sizes