According to Sina Finance, AMD and Google are gaining ground in the AI chip market as Nvidia customers explore alternatives, a shift that may accelerate after Jensen Huang stepped down as Nvidia CEO last week. Nvidia’s board named CFO Colette Kress interim CEO. Reports of multiple extramarital affairs preceded Huang’s departure.
AMD chips had already attracted interest before the leadership change. Cirrascale CEO David Driggers said customers had asked since last year about using AMD chips for AI workloads, particularly inference. Tensorwave COO Piotr Tomasik said winning customer trust had been difficult, but external benchmarks last year showed AMD chips matching or outperforming Nvidia in some tests. OpenAI, Anthropic and Meta said they would use AMD chips from the second half of 2025, and OpenAI announced this week that it would use Tensorwave’s cloud services.
Nvidia remains far larger by revenue: its fiscal third-quarter revenue was $115.8 billion, more than eight times AMD’s third-quarter revenue, reported on November 3. But some customers are drawn to AMD’s lower prices and improved ROCm software. Tensorwave said staff now spend 80% to 90% less time helping customers resolve ROCm problems than they did last year. The company expects annual recurring revenue of $210 million by the end of 2026, compared with $125 million at the end of 2025. Sciforium said its customer base has tripled since its first commercial contract in June, computing capacity has grown fivefold over the past year, and its team has nearly doubled; it is raising funds to expand capacity.
Google is also expanding its TPU business. In July, it disclosed guarantees of up to $44 billion for data-center leases. Cirrascale said customers had expressed interest in spending hundreds of millions of dollars on computing services using Google’s TPU 8 chips. Autolane has moved some workloads from GPUs to TPUs since September, citing lower costs, though it still relies heavily on GPUs. DoiT CEO Vadim Solovey said 16 of the company’s roughly 5,000 customers use TPUs at scale, describing them as a second supplier rather than a replacement.
Nvidia retains advantages including access to TSMC manufacturing capacity and a strong sales operation. Customers cited longer waits for AMD chips than for Nvidia products, and Driggers said customers continued to increase Nvidia orders even as its market share slipped.