OpenAI Teams Up with AMD in Landmark Deal to Expand AI Computing Power

Nvidia’s market dominance will be challenged by the five-year agreement as OpenAI prepares to implement AMD’s new MI450 chips. OpenAI CEO Sam Altman and Advanced Micro Devices CEO Lisa Su during a hearing of the Senate Commerce, Science, and Transportation Committee in Washington in May.

In one of the most direct challenges to industry leader Nvidia to date, OpenAI and chip designer Advanced Micro Devices announced a multibillion-dollar collaboration to work together on AI data centres that will run on AMD processors. As per the agreement, OpenAI agreed to buy six gigawatts of AMD CPUs, beginning with the MI450 chip the following year.

The manufacturer of ChatGPT will purchase the chips directly or via its partners in cloud computing. In an interview on Sunday, Lisa Su, the head of AMD, stated that the deal will generate tens of billions of dollars in new revenue for the chip company over the next five years. AMD stated that the proposal costs tens of billions of dollars per gigawatt of computing capacity, while the two companies did not reveal the estimated total cost.

AMD’s Biggest Deal Till Now

The agreement is AMD’s largest victory in its attempt to challenge Nvidia’s hegemony in the AI chip market. AMD hasn’t had much of an impact on the rapidly expanding market for the more expensive supercomputing chips required by sophisticated AI systems, despite the fact that its processors are extensively used for gaming, in personal computers, and in conventional data centre servers. Inference functions—the calculations that enable AI applications like chatbots to reply to user enquiries—will be performed by OpenAI using AMD CPUs.

In a joint interview with Su, OpenAI Chief Executive Sam Altman stated that the demand for inference computing has increased dramatically as the abundance of huge language models and other tools has increased. In a joint interview with Su, Altman stated, “It’s hard to overstate how difficult it’s become” to get enough computing power. “It takes some time, but we want it rapidly, he added.

According to the two CEOs, the agreement will unite their businesses and provide them with incentives to invest in the AI infrastructure growth. Both businesses benefit, according to Su, who added, “I’m glad that OpenAI’s incentives are tied to AMD’s success and vice versa.”

Nvidia to Face Tougher Challenges in Future

Although Nvidia is still the chip of choice for AI businesses, it is up against competition from nearly every industry segment. AI chips are designed and sold by cloud giants like Google and Amazon, while OpenAI recently inked a $10 billion agreement with Broadcom to develop its own proprietary chip.

Next year, Nvidia will release its eagerly awaited Vera Rubin processor, which is expected to be more than twice as powerful as its current iteration, the Grace Blackwell. In the second half of next year, OpenAI will start running its AI models on a 1-gigawatt MI450 chip.

Altman stated that as demand for AI services—as well as the associated processing and infrastructure requirements—is expected to greatly exceed supply, the future of many businesses will become more intertwined.