OpenAI announced a strategic partnership with AMD under which the AI firm will purchase tens of billions of dollars’ worth of AMD’s next-generation GPUs, starting with deployments of 1 gigawatt in 2026 and scaling to 6 gigawatts. AMD has issued warrants giving OpenAI the option to acquire up to 160 million shares—roughly 10 percent of the company—if certain deployment and share price milestones are met. AMD expects the new deal to generate “tens of billions” in new revenue over time, and the news sent AMD’s stock soaring by over 30 percent in early trading.
Key Takeaways
– OpenAI will acquire AMD GPUs across multiple future generations, starting with the MI450 series, in a structured deployment plan up to 6 gigawatts overall.
– The partnership includes a warrant agreement for OpenAI to buy up to 160 million shares of AMD—about 10%—if technical, commercial, and share-price milestones are met.
– The deal is being viewed as a transformative move for both firms: it accelerates AMD’s revenue growth in AI infrastructure and helps OpenAI diversify its hardware supply chain beyond reliance on Nvidia.
In-Depth
The new agreement between OpenAI and AMD marks a bold move in the high-stakes race to build next-generation AI infrastructure. Under the terms, AMD will supply large volumes of GPUs to OpenAI over multiple years, starting with a 1 gigawatt deployment in the second half of 2026. Over time, the total commitment is for 6 gigawatts of compute capacity. In tandem, AMD has granted OpenAI a warrant to acquire up to 160 million shares of AMD common stock. If all vesting conditions are met, that could represent nearly a 10 percent stake in AMD. The vesting of the first tranche is tied to the initial deployment milestone, and further tranches vest as OpenAI scales purchases and AMD hits certain share-price thresholds.
For AMD, the deal offers a strong endorsement of its AI chip roadmap and opens opportunities for “tens of billions” in incremental revenue. AMD anticipates that the partnership will unlock additional downstream demand from customers who see AMD validated as a major AI compute supplier. On OpenAI’s side, this deal helps reduce dependence on Nvidia by diversifying its supplier base. It gives OpenAI more leverage and optionality in matching hardware to evolving AI workloads.
Of course, the structure of the warrant agreement imposes rigorous performance and market conditions. If AMD’s stock fails to meet the thresholds, or OpenAI doesn’t scale deployments quickly enough, not all of the warrant shares will vest. But the alignment is clever: both parties’ incentives are tied together. In effect, OpenAI becomes not just a buyer but a contingent investor in AMD’s success.
The market response was dramatic. AMD’s stock surged more than 30 percent on the announcement, adding tens of billions to its market value in a single session. That kind of reaction highlights how deeply investors sense the strategic implications: GPUs and AI compute are now among the most critical battlegrounds in tech.
Even though Nvidia remains a dominant force in the AI hardware domain, this deal signals that AMD is being taken seriously as an alternative. For OpenAI, the move helps guard against supply chain risk and gives it additional bargaining power as it scales its data center footprint.
At the same time, the deal underscores how capital-intensive AI infrastructure is becoming. Deploying gigawatt-scale compute requires not just chip procurement but massive power, cooling, and facility investment. OpenAI must hit its milestones not only in purchasing hardware but in building ecosystems to make that hardware useful in production. The success of this deal will depend on both hardware engineering and systems integration over coming years.
In short, the OpenAI-AMD pact is a bold bet by both sides: AMD to gain a major role in AI compute, and OpenAI to hedge its bets and embed itself more deeply into the hardware layer of AI infrastructure.