According to CNBC, Google's AI infrastructure chief, Amin Vahdat, told an all-hands meeting on November 6 that the company must dramatically scale its compute capacity: doubling it every six months and achieving an additional ~1,000 times over the next four to five years to meet the exploding demand for AI.

Vahdat stressed that AI infrastructure is the most critical and expensive part of the race, but Google’s goal is to build the world’s most reliable, performant, and scalable system—not just outspend rivals.

CEO Sundar Pichai, also present, warned that the 2026 AI competition will be “very intense” and dismissed bubble fears: “The risk is under-investment, not over-investment.” He noted that Google Cloud grew 34% to over $15B this quarter, with a $155B backlog, figures that would be even higher with more compute available.

Google is also boosting efficiency through new models and its 7th-generation TPUs, which are ~30 times more energy-efficient than the original 2018 model.

The above content is compiled by ModeZone, a fashion and entertainment magazine.