Greetings,

Nvidia's $4 trillion dominance in the AI chip market has seemed unshakable, but a major rival is changing tactics to encroach on that turf. Amir Efrati, Erin Woo, and Anissa Gardizy report that Google is picking up the pace of its efforts to compete directly with Nvidia by pitching customers on using Google's tensor processing units (TPUs) in their own data centers.

For years, the search giant mostly rented these chips to cloud customers, but now it is actively pitching giants like Meta Platforms and financial institutions on the idea of on-premise use. It represents a significant strategic shift to capture more of the market.

Why it caught my eye:

• A Multi-Billion Dollar Shift: Meta is currently in talks with Google about spending billions of dollars to use TPUs in its own data centers by 2027, a move designed to diversify away from Nvidia.
• Closing the Technical Gap: While analysts often assume Nvidia is untouchable for training large models, Meta has discussed using Google's TPUs for training new AI models, suggesting Google has narrowed the lead.
• Aggressive Revenue Targets: Google is playing for keeps—leaders in its cloud unit have suggested this push could help them grab as much as 10% of Nvidia's annual revenue.

The full story dives deeper into the specific software Google is building to chip away at Nvidia's moat and how Jensen Huang is responding. It's a must read on the future of AI infrastructure.

Best,

Jessica Lessin
Founder & Editor-in-Chief

Google Further Encroaches on Nvidia's Turf With New AI Chip Push

Google is picking up the pace of its efforts to compete directly with Nvidia in the AI chip business.

For years, the search giant has rented its own AI chips, known as tensor processing units, to cloud customers who use them in its Google Cloud data centers. Now, though, Google has begun pitching some of those customers—including Meta Platforms and big financial institutions—on the idea of using TPUs in their own data centers, according to people involved in the talks.