Google Has Two New Weapons in Its Fight Against Nvidia
What happened
Google introduced two new versions of its AI chips, named TPU 8tx and TPU 8i Max, on Monday. These chips are designed to address secure data processing needs and support high-memory AI workloads. The TPU 8tx targets government and regulated industries requiring advanced encryption, while the TPU 8i Max is built for tasks needing large memory and computing power. Both chip systems are scheduled to become available in 2027.
Key details
- TPU 8tx uses hardware-level security for data protection during processing.
- TPU 8i Max combines two TPU 8i chips, doubling computing power and memory bandwidth.
- Both chips are expected to be available in 2027.
- Google Cloud's revenue from TPU systems more than doubled year over year to $30.7 billion for the quarter ending September 30.
- Google has nearly 100 TPU customers, including Apple, Meta, and Oracle.
Why it matters
Businesses in regulated sectors and those running complex AI workloads will have new options for secure and high-memory AI processing from Google starting in 2027. This gives companies alternatives to Nvidia's offerings, especially for sensitive data and large-scale AI operations. Organizations using Google Cloud or seeking to build their own AI infrastructure may see increased capabilities and potentially more competitive pricing or features.