AMD Says Its Processors Power Google Cloud TPU 8tx Infrastructure Ahead of Share Decline
Key Facts
AMD announced on November 9 that its fifth-generation EPYC processors power the host infrastructure for Google Cloud TPU 8tx, expanding its presence in cloud infrastructure for artificial intelligence. The company said the service is the first cloud TPU system to offer hardware-enforced confidential computing. The announcement came four days before AMD shares closed at $534.06 on November 13, compared with $508.58 at the close on November 9. In the same session, Nvidia shares ended trading at $162.85, after closing at $178.67 on November 9.
AMD describes the role of EPYC processors in TPU 8tx systems as extending beyond providing host computing capacity: they coordinate workloads, prepare data and provide high-throughput connectivity with TPU accelerators. This places the processors in an operating layer that links system resources and supports data flows to the accelerators. The company said AMD SEV technologies are used to provide confidential computing. It describes the service as the first cloud TPU system to enable this type of computing through hardware-enforced measures. The announcement concerns Google Cloud infrastructure and the TPU 8tx service, and by itself does not establish the scale of revenue or commercial demand resulting from it.
AMD data show that data centers now account for a larger share of the company’s revenue; the segment contributed 64% of total revenue in the third quarter of 2026, up from 47% in the same period a year earlier. The company said the segment’s revenue growth accelerated quarter over quarter in each quarter of the year. These figures link the expansion of data center activity with the importance of server processors to AMD’s business. The Google Cloud announcement also adds a specific example of EPYC being used in host infrastructure for artificial intelligence systems. The available facts here do not include total quarterly revenue figures or a breakdown of the value contributed by this collaboration.
The divergence between the two stocks came in a week that saw a management development at Nvidia: the company disclosed in a filing with the U.S. Securities and Exchange Commission that its board accepted Jensen Huang’s resignation as chairman, chief executive and a board member on November 11. The filing said Chief Financial Officer Colette Kress would serve as interim chief executive while a successor was sought. On November 12, Nvidia shares closed down 4.3%, according to Reuters, while AMD and Broadcom shares each rose about 6%. Reuters reported that Nvidia shares recovered 1% in after-hours trading. The move shows the market’s reaction in the November 12 session, but by itself is not enough to attribute the full price change between November 9 and 13 to the resignation.
In Friday’s session on November 13, Nvidia continued to fall, according to the Associated Press, while the Nasdaq index declined 0.9%. AMD shares closed at $534.06, compared with $508.58 on November 9, while Nvidia shares ended the session at $162.85, compared with $178.67 on the earlier date. The closing prices reflect opposing trajectories between the two sessions, with the broader index still under pressure on Friday. The available facts do not include enough figures to calculate a consistent five-trading-day performance difference or verify its historical comparison since 1999. Nor do they include a reliable closing price for Google shares on November 13.