Google’s AI infra spend triples to over $90 billion as Pichai plays down bubble fears

Google was previously spending less than $30 billion per year on artificial intelligence (AI) infrastructure. That figure has now risen to more than $90 billion, said Google chief executive Sundar Pichai in an interview with BBC Newsnight a week ago.

He added that if similar investments across the industry are combined, the total would surpass one trillion dollars.

Pichai also outlined aggressive scaling plans, saying the company will build in “just a couple of years” what previously took 10 to 20 years to develop.


On the AI bubble fears

Addressing concerns about a potential AI bubble, Pichai said the focus should be on the “actual progress” Google has made in advancing its model capabilities, describing the developments as “palpably exciting.” He added that consumers are increasingly adopting and using Google’s AI-powered products.

“We see real demand, and we are constrained in our ability to serve that demand,” he said.

Drawing a comparison with the early internet era, Pichai noted that heavy investments during that period were not questioned and argued that the same perspective should apply to AI. He said internet investments ultimately “fundamentally changed how we work digitally as a society.”

On accuracy and AI regulation

When asked about errors in AI-generated responses, Pichai said Google Search remains the primary tool for delivering accurate information within Gemini.

He acknowledged the limitations of current state-of-the-art AI systems, adding that solutions should focus on strengthening and enriching information ecosystems.