In a deal defined by its staggering scale, OpenAI and AMD have entered into a partnership that will channel an unprecedented amount of energy and computing power into the development of artificial intelligence. The agreement, valued in the multibillions, will provide OpenAI with six gigawatts of computing capacity, a figure that highlights the voracious energy demands of the modern AI arms race.
This immense power commitment, equivalent to the annual output of the Hoover Dam three times over, will be realized through the deployment of hundreds of thousands of AMD’s next-generation AI chips. The hardware rollout is scheduled to begin in the latter half of 2026, with plans for OpenAI to construct a one-gigawatt facility based on AMD’s upcoming MI450 series of chips as early as next year.
The partnership represents a critical infrastructure investment for OpenAI, whose CEO, Sam Altman, has identified the scarcity of computing power as the primary obstacle to progress. By securing this massive supply of GPUs from AMD, OpenAI is building the foundation needed to train AI models of unprecedented complexity and capability, pushing the boundaries of what machines can learn and create.
Beyond the hardware, the deal establishes a powerful financial link between the two tech giants. AMD has issued a warrant giving OpenAI the right to purchase a 10% stake in the company, amounting to 160 million shares, for a mere one cent per share. This arrangement ensures that as AMD’s technology powers OpenAI’s growth, OpenAI will also share in the chipmaker’s resulting financial success.
The announcement sent shockwaves through the financial markets, causing AMD’s valuation to jump by $80 billion in a single day. AMD leadership believes this is just the beginning, projecting that the influence of this “pioneer” deal will drive over $100 billion in new revenue over the next four years as other companies follow OpenAI’s lead.
