AMD Secures Major AI Chip Deal with OpenAI

• AMD lands multi-year supply agreement worth tens of billions annually, with OpenAI gaining option to acquire up to 10% stake in the chipmaker.

AMD announced Monday a significant multi-year agreement to supply artificial intelligence chips to OpenAI, a deal expected to generate tens of billions of dollars in annual revenue for the semiconductor company. The arrangement includes a warrant allowing OpenAI to purchase up to 10% of AMD’s shares, marking a substantial commitment from the ChatGPT creator to diversify its chip supply chain. AMD’s stock surged more than 34% following the announcement, reaching $226.71—a one-year high—and adding approximately $80 billion to the company’s market capitalization. The agreement represents one of the most substantial partnerships in the AI hardware sector, underscoring the industry’s intense demand for computing infrastructure.

Massive Scale Deployment Starting 2026

The deal encompasses the deployment of hundreds of thousands of AMD graphics processing units (GPUs) equivalent to six gigawatts of computing power over several years, beginning in the second half of 2026. This energy requirement roughly equals the consumption of 5 million U.S. households or three times the electricity output of the Hoover Dam. OpenAI plans to construct a one-gigawatt facility using AMD’s forthcoming MI450 series chips starting next year, with AMD expecting to recognize revenue at that point. The phased rollout reflects the enormous infrastructure investments required to support advanced AI development and training workloads.

AMD executives project the agreement will generate more than $100 billion in new revenue over four years when accounting for ripple effects across the broader customer base. “Other people are going to come along with it because this is really the pioneer, a pioneer in the industry that has a lot of influence over the broader ecosystem,” explained AMD strategy chief Mat Hein. Forrest Norrod, AMD’s executive vice president, characterized the arrangement as “transformative, not just for AMD, but for the dynamics of the industry.” Analysts currently estimate AMD’s total 2025 revenue at $32.78 billion, suggesting this OpenAI partnership could significantly reshape the company’s financial trajectory.

Warrant Structure and Strategic Investment

Under the agreement’s terms, AMD issued a warrant granting OpenAI the right to purchase up to 160 million AMD shares at 1 cent each throughout the duration of the chip supply deal. The warrant vests in tranches tied to specific milestones agreed upon by both companies, including the initial MI450 chip shipment scheduled for late 2026. Subsequent vesting milestones include AMD stock price targets that escalate to $600 per share for the final tranche to unlock. This structure aligns OpenAI’s potential equity stake with AMD’s performance and the successful execution of the chip deployment, creating mutual incentives for both parties.

OpenAI CEO Sam Altman stated the AMD partnership will help the company build sufficient AI infrastructure to meet its requirements. However, the funding mechanism for this massive commitment remains unclear, particularly given OpenAI’s current financial position. Media reports indicate OpenAI generated approximately $4.3 billion in revenue during the first half of 2025 while burning through $2.5 billion in cash, as the startup invests heavily in talent acquisition and AI tool development. OpenAI currently holds a $500 billion valuation following its recent funding rounds.

Diversification Beyond Nvidia Dominance

OpenAI has maintained a relationship with AMD for years, previously providing design input for earlier chip generations including the MI300X series. The company has been actively securing multiple chip sources to support its ambitious expansion plans, having announced a separate agreement with Nvidia in September for up to $100 billion in investments. That arrangement includes supplying at least 10 gigawatts of Nvidia systems, with OpenAI deploying one gigawatt of Nvidia’s next-generation Vera Rubin chips in late 2026. Reports from The Information suggest Altman has set expectations of reaching 250 gigawatts of total compute capacity by 2033.

Beyond external suppliers, OpenAI is developing its own custom silicon for AI applications through a partnership with Broadcom, as Reuters previously reported. Major cloud computing providers like Google and Amazon have similarly invested in proprietary processor development to reduce dependency on third-party chip manufacturers. OpenAI and its primary backer Microsoft also announced last month a non-binding agreement to restructure OpenAI into a for-profit entity, signaling continued governance evolution. According to sources familiar with the matter, the AMD deal does not alter OpenAI’s existing compute plans or its partnership arrangements with Microsoft.