Google’s AI infrastructure chief told employees the company must double its compute capacity every six months to support rising demand for artificial intelligence services. At a recent all-hands meeting, Google Cloud vice president Amin Vahdat described compute growth as essential to staying competitive, noting that the next wave of AI demand could require a thousandfold increase in capability within five years.
The remarks came shortly after Alphabet raised its capital expenditures forecast for the second time this year, lifting planned spending to as much as $93 billion. Hyperscaler rivals Microsoft, Amazon and Meta have issued similar boosts, with the group projected to collectively invest more than $380 billion this year as they race to expand data center capacity and train larger models.
Vahdat said Google is focused on building infrastructure that is more reliable and efficient rather than simply outspending competitors. The company highlighted improvements from custom silicon, including Ironwood, its seventh-generation Tensor Processing Unit, which it says is nearly 30 times more power efficient than its first TPU released in 2018.
CEO Sundar Pichai told employees that 2026 will be an intense period as cloud demand grows and competition tightens. He acknowledged broader concerns about a potential AI bubble but emphasized the risk of underinvesting, pointing to strong growth in Google Cloud’s $155 billion backlog as evidence of sustained demand.
The comments reflect ongoing debate across the industry about long-term AI spending, a topic that has surfaced in recent discussions of whether the sector may be entering a bubble phase.
