Google CEO Sundar Pichai says company is giving up short-term cloud revenue to prioritise advanced AI development
Google CEO Sundar Pichai has said the company is giving up potential short-term revenue from its Cloud business to ensure its researchers have enough computing power to develop advanced artificial intelligence models.
The decision highlights the growing pressure on technology companies to balance commercial demand with the resources needed to compete in the AI race.
Speaking to Alex Kantrowitz on the Big Technology Podcast, Pichai explained that Google was making deliberate choices about how to distribute its limited computing resources. Rather than directing all available capacity towards paying Cloud customers, the company is reserving some of it for its AI research teams.
"Our cloud business is at any given moment in time foregoing short-term revenues because we are making compute trade-offs to support frontier AI development," Pichai said during the interview.
The remarks offer a clearer picture of Google's priorities as competition intensifies among companies developing increasingly capable AI models. They also reveal the practical challenges behind that ambition, particularly when computing infrastructure cannot be expanded quickly enough to meet every demand.
Pichai described providing sufficient computing capacity for Google DeepMind, the company's AI research division, as a non-negotiable priority. He said ensuring the team has the resources required to develop cutting-edge models remains the starting point for decisions about computing capacity across the business.
"Ground zero for us is making sure we are investing in the compute that's needed to be at the AI frontier," he said.
Computing power is essential to developing and running advanced AI systems. It depends on access to specialised chips, servers and data centres, all of which require significant investment and time to build.
Google's Cloud business, meanwhile, sells computing capacity and related services to businesses that need infrastructure for their own operations. Allocating more resources to internal AI research can therefore mean having less capacity available to meet external demand.
Pichai said Google's different businesses still prepare their own long-term plans based on projected demand, customer requirements and existing contracts. However, he made it clear that keeping DeepMind equipped to work at the frontier of AI development takes precedence.
The strategy reflects Google's view that staying competitive in advanced AI is important across its wider business, including Search and Cloud. Pichai said developing leading models also gives researchers an earlier understanding of emerging capabilities, which can help the company build future products.
The challenge is not simply deciding where to direct computing power. Google must also plan infrastructure years in advance, making it difficult to respond quickly when demand rises unexpectedly.
Pichai pointed to the long timelines involved in building data centres and securing the necessary computing equipment. Explaining why today's capacity reflects decisions made well before the current demand became clear, he said, "The compute we have today is a function of decisions I made in 2024."
He added that the shortage was not limited to Google. Technology companies across the industry are struggling to secure as much computing capacity as they would like, even as demand for AI services grows.
The pressure is particularly visible in requests for future infrastructure. Pichai said accommodating a commercial customer seeking computing capacity in 2027 would be "incredibly difficult to do". He also noted that Google had leased some short-term computing capacity in the previous year to help meet customer requirements.
These constraints create a difficult balancing act. Google must plan for future customer demand while ensuring its research teams have the resources to develop more advanced models. Decisions made today could also prove less effective than expected if AI technology and demand change significantly over the next few years.
For Google, however, Pichai's comments suggest that the immediate priority remains clear. The company is prepared to forgo some short-term Cloud revenue if that helps maintain its position in advanced AI development.
Trisha Bhattacharya is a Senior Content Producer at Livemint, with over two years of experience covering entertainment news from India and beyond. She spends her days tracking what’s trending, breaking down pop culture moments, and turning fast-moving entertainment stories into sharp, engaging reads that actually make people want to click — and stay. <br> She holds a Master’s degree in English Literature from Lucknow University, a background that shapes her love for layered narratives, strong voices, and stories that linger long after they’re told. Before joining Livemint, Trisha worked with India Today as an entertainment journalist and film critic. There, she reviewed films, covered industry news, and built a strong foundation in storytelling and cultural analysis. <br> Trisha enjoys working at the intersection of media, culture, and audience interest, always looking for fresh angles and formats. Films, shows, and music are not just her beat but her biggest passion — something that naturally reflects in her writing. Whether it’s cinema, streaming shows, music, or internet trends, she approaches every story with curiosity and intent. <br> Outside the job description, she’s unapologetically passionate about films, shows, and music — sometimes a little too passionate, if you ask her. That enthusiasm often spills into her work, adding personality, urgency, and a touch of chaos that keeps her writing alive. For Trisha, entertainment isn’t just a beat — it’s a language she speaks fluently.
Catch all the Business News , Corporate news , Breaking News Events and Latest News Updates on Live Mint. Download The Mint News App to get Daily Market Updates.
More
✕
Wait for it…
Oops! Looks like you have exceeded the limit to bookmark the image. Remove some to bookmark this image.