US chipmaker AMD said Wednesday that it is set to invest up to $5 billion in artificial intelligence start-up Anthropic in a partnership that also includes a major order for processors. The deal signals the ambition of the California-based AMD as it seeks to position itself as a competitor to Nvidia in the market for high-performance AI chips. Anthropic is set to purchase up to two gigawatts of AMD's latest-generation chips, known as the Instinct MI450, under the agreement. Anthropic will also deploy AMD's Helios rack-scale solutions, another system AMD is developing to compete with Nvidia in the AI buildout. Deployment of the first gigawatt is due to begin in 2027, AMD said. "This collaboration brings together Anthropic's leadership in frontier AI with the full strength of AMD high-performance computing," said AMD chief executive Lisa Su, noting that both companies look to speed up AI adoption at scale. "By partnering with AMD across the stack, we are securing the capacity we need and optimizing it for training and serving Claude," said Anthropic chief compute officer Tom Brown. The partnership comes on the heels of a multi-year AMD venture announced in October with OpenAI to develop AI data centers. While Nvidia's graphics processing units (GPUs) are by far the dominant player in the AI field, AMD's ventures with emerging AI leaders OpenAI and now Anthropic help the company to burnish its credentials. However, some investors question whether the orders are artificially inflating chip demand. Under the October deal, AMD granted to OpenAI a warrant for up to 160 million shares of AMD -- a financial product that can be converted to shares under certain conditions, leaving the AI giant in a position to own roughly 10 percent of AMD. AMD committed Wednesday to making "a strategic equity investment" of up to $5 billion in Anthropic, representing less than one percent of Anthropic's estimated $965 billion valuation. Shares of AMD rose 1.0 percent in morning trading.
