OpenAI and Advanced Micro Devices (AMD) have announced a landmark, multi-year strategic partnership under which AMD will supply AI chips to OpenAI and grant the AI firm an option to acquire up to 10% of AMD through a structured warrant agreement.

Under the agreement, OpenAI will deploy as much as 6 gigawatts of GPUs over several years, beginning with an initial 1 gigawatt deployment in the second half of 2026, using AMD’s upcoming MI450 series. The scale of this commitment underscores the intensifying arms race in AI compute infrastructure.

In parallel, OpenAI receives a warrant to purchase 160 million AMD shares at $0.01 per share, contingent on achieving milestones tied to chip deployment and AMD’s share price reaching specified targets. The stake option effectively allows OpenAI to become a significant investor in AMD if the partnership delivers results.

AMD’s leadership cast the deal as transformative. The company expects significant new revenue streams over the next several years from this arrangement, projecting that the deal could generate tens of billions annually. The bullish outlook comes amid growing ambition by AI firms to diversify beyond dominant incumbents like Nvidia.

Markets reacted swiftly: AMD’s stock jumped over 23% in premarket trading, hitting highs not seen since March 2024.

For OpenAI, the partnership addresses a persistent challenge — securing reliable, scalable compute capacity. While the firm has historically leaned on Nvidia systems and cloud providers, this deal reflects a pivot toward deeper hardware alignment and self-sufficiency.

This move is likely to ripple across the AI ecosystem. Competitors will scrutinize the terms; chipmakers may recalibrate strategy; and the alignment of hardware-software under one roof could accelerate innovation cycles.

The next few years will test execution, integration, and the ability to deliver at scale. If successful, this deal may redefine how power, performance, and ownership interplay in the AI era.