Google AI Researchers Compete for Internal Compute as Demand Surges Across $2 Trillion AI Race
Updated
Updated · Bloomberg · May 18
Google AI Researchers Compete for Internal Compute as Demand Surges Across $2 Trillion AI Race
6 articles · Updated · Bloomberg · May 18
Google’s own AI researchers are competing internally for access to the company’s computing capacity, underscoring how scarce high-end AI infrastructure has become even inside a top provider.
That squeeze reflects heavy demand for resources Google has built into a strategic advantage: a sizable cloud business, in-house chip development and agreements to share capacity with companies including Anthropic and Meta.
The bottleneck shows how the AI buildout is shifting from model development alone to control of chips and data-center compute, with access to infrastructure becoming a key constraint across the industry.
By renting its best AI chips to rivals, is Google accelerating its own obsolescence?
Google owns the world's most powerful AI infrastructure. Why are its own developers falling behind?