Google Plans a Thousandfold Increase in AI Capacity

Google must double its computing power every six months to meet AI demand.

Google is compelled to double its computing power every six months to meet the growing demand for artificial intelligence technologies. This was revealed by Google Cloud Vice President Amin Vahdat during a meeting with employees, as reported by CNBC.

According to the presentation, over the next 4–5 years, capacity must double every six months, leading to an increase of approximately 1000 times.

“Competition in AI infrastructure is the most important and at the same time the most expensive part of the race in artificial intelligence,” said Vahdat.

The presentation took place on November 6, a week after the release of the third-quarter financial report. The company raised its capital expenditure forecast for the second time this year to $91-93 billion and announced plans to “significantly increase” it in 2026.

Microsoft, Amazon, and Meta also plan to “significantly increase” their spending. Collectively, the four corporations plan to spend $380 billion in 2025.

“Google’s task is to build this infrastructure. But it doesn’t necessarily have to spend more than its competitors,” Vahdat stated.

The corporation aims to provide a “more reliable, productive, and scalable foundation” compared to its competitors.

Vahdat mentioned that Google is increasing capacity through more efficient models and its own chips.

Last week, the company announced the launch of the seventh generation of Tensor Processing Unit processors called Ironwood. They are 30 times more energy-efficient compared to the first version of Cloud TPU from 2018.

The Google Cloud Vice President noted the company’s competitive advantage thanks to its research division, DeepMind.

“Google needs to ensure 1000 times more power, computing resources, and network storage at almost the same cost and energy consumption. It won’t be easy, but through collaboration and joint design, we will achieve our goals,” Vahdat said.

High Competition and a Bubble

Google CEO Sundar Pichai told employees that 2026 will be challenging due to growing competition in artificial intelligence and high demand for cloud and computing resources.

One participant asked whether a bubble is forming in the industry that could burst and slow down further development.

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

He returned to the topic of risks from overly conservative investments and highlighted Google’s cloud business, which recorded a quarterly revenue growth of 34% to over $15 billion. The order book for this segment reached $155 billion.

“We are better prepared than others to withstand setbacks,” Pichai emphasized.

Nvidia Cools the Hype

Talk of a bubble intensified ahead of Nvidia’s quarterly report. Shares of major AI players fell significantly.

A strong report from a player with a market capitalization of over $4.5 trillion and a roughly 8% share in the S&P 500 index led to a rise in shares of the American chipmaker and other tech companies.

In November, Pichai said that there are “elements of irrationality” in the market, and if a bubble bursts, “no company will be spared.”

Analysts and chip manufacturers warned of a memory chip shortage that could affect consumer electronics and the automotive industry in 2026.