Google stockpiling TPUs for AGI development while expanding third-party compute

A major cloud provider is prioritizing internal AI research over customer capacity, signaling strategic shift in compute allocation.

Abstract illustration of computing resources divided between internal and external allocation
AI-generated illustration · Sylvaris

Internal research takes priority over customer workloads

Google is reserving substantial TPU capacity for artificial general intelligence research rather than allocating it to Google Cloud customers. The company is simultaneously purchasing third-party compute capacity to meet external demand.

The approach marks a shift from previous infrastructure allocation strategies where cloud customer needs typically received priority. Google has not disclosed specific TPU quantities being reserved or the timeline for AGI development efforts.

Third-party capacity addresses customer demand

Google Cloud is expanding partnerships with external compute providers to handle customer workloads typically served by TPUs. This includes arrangements with GPU-focused infrastructure companies and specialized AI cloud providers.

The strategy allows Google to maintain research momentum while preserving cloud business commitments. TPU architecture remains proprietary to Google, making the custom silicon particularly valuable for internal AI development work.

sources
more in Cloud & Infrastructure
Google Cloud becomes Alphabet's fastest-growing segment at over 20% of revenue Cloud infrastructure now represents more than one-fifth of Alphabet's total revenue and operating profit, marking a strategic shift beyond advertising dominance. US Army exhausts annual AI token allocation, implements usage limits Rapid AI adoption in military operations faces practical constraints when organizations underestimate infrastructure requirements. AMD and Anthropic reach $5 billion AI infrastructure deal AMD's largest AI infrastructure commitment to date signals intensifying competition in the data center GPU market dominated by NVIDIA.