Google: AI Compute Demand Requires Doubling Every 6 Months

Google’s AI infrastructure division faces a monumental task: doubling its compute capacity every six months to keep pace with the surging demand for artificial intelligence services, according to Amin Vahdat, a Google Cloud VP.

During an all-hands meeting on Nov. 6, Vahdat presented an “AI Infrastructure” overview, revealing a slide titled “AI compute demand.” The slide starkly stated, “Now we must double every 6 months…. the next 1000x in 4-5 years.”

“The competition in AI infrastructure is the most critical and also the most expensive part of the AI race,” Vahdat emphasized during the meeting, which also featured Alphabet CEO Sundar Pichai and CFO Anat Ashkenazi fielding employee questions.

This presentation came a week after Alphabet reported stronger-than-anticipated third-quarter earnings, leading to a second upward revision of its capital expenditures forecast for the year, now projected between $91 billion and $93 billion. Executives have also signaled a “significant increase” in capex for 2026. This echoes similar moves by hyperscaler rivals Microsoft, Amazon, and Meta, with the collective spending of these four tech giants now poised to exceed $380 billion this year. The massive investments underscore the escalating infrastructure arms race in the AI domain, as each company strives to establish a dominant position.

Vahdat acknowledged the capital intensity of the AI race, stating, that Google’s first priority “is of course to build this infrastructure but it’s not to outspend the competition, necessarily,” Vahdat said. “We’re going to spend a lot,” he said, adding that the real goal is to provide infrastructure that is far “more reliable, more performant and more scalable than what’s available anywhere else.”

Beyond infrastructure expansion, Google is also focusing on model efficiency and custom silicon to bolster capacity. The company recently announced the public launch of Ironwood, its seventh-generation Tensor Processing Unit (TPU), claiming nearly 30 times greater power efficiency compared to its first Cloud TPU from 2018. This focus on custom silicon is critical. Google’s TPUs are designed specifically for machine learning workloads, potentially offering a performance advantage over general-purpose processors. Furthermore, enhancing power efficiency not only reduces operational costs but also aligns with growing concerns about the environmental impact of large-scale compute infrastructure.

Vahdat highlighted DeepMind as a key asset, leveraging its research to anticipate future AI model architectures. This forward-looking approach informs Google’s infrastructure planning, ensuring it stays ahead of the curve as AI technology rapidly evolves.

Google aims to “be able to deliver 1,000 times more capability, compute, storage networking for essentially the same cost and increasingly, the same power, the same energy level,” Vahdat stated. “It won’t be easy but through collaboration and co-design, we’re going to get there.” This ambitious goal highlights the need for innovation in hardware, software, and system architecture to drive down the cost per unit of compute, a crucial factor for sustainable growth in the AI era.

Pichai cautioned employees that 2026 will be “intense,” citing fierce AI competition and the immense pressure to meet escalating cloud and compute demands.

He also addressed the growing concerns about a potential AI bubble, a topic gaining traction in Silicon Valley and on Wall Street as some investors question the justification for the massive spending expected in the coming years.

An employee question that Pichai read aloud posed: “Amid significant Al investments and market talk of a potential Al bubble burst, how are we thinking about ensuring long-term sustainability and profitability if the Al market doesn’t mature as expected?”

Pichai acknowledged the validity of these concerns.

“It’s a great question. It’s been definitely in the zeitgeist, people are talking about it,” Pichai said.

He reiterated his past warnings about the risks of underinvesting in the burgeoning AI sector and highlighted the robust performance of Google’s cloud business, which posted a 34% year-over-year revenue increase to over $15 billion in the last quarter, with a backlog reaching $155 billion. This strong cloud performance provides a critical foundation for funding AI ambitions.

“I think it’s always difficult during these moments because the risk of underinvesting is pretty high,” Pichai said. “I actually think for how extraordinary the cloud numbers were, those numbers would have been much better if we had more compute.” This statement underscores the direct link between compute capacity and revenue generation in the AI-driven cloud market.

He emphasized Google’s disciplined approach, citing the strength of its underlying businesses and its solid balance sheet. This suggests Google is confident in its ability to weather any potential market downturn.

“We are better positioned to withstand, you know, misses, than other companies,” Pichai said.

Looking ahead to the coming year, Pichai warned employees of inevitable “ups and downs.”

“It’s a very competitive moment so, you can’t rest on your laurels,” he said. “We have a lot of hard work ahead but again, I think we are well positioned through this moment.”

Google declined to comment.