OpenAI taps Google Cloud to power ChatGPT as demand surges, reducing reliance on Microsoft: Report

OpenAI has entered into a partnership with Google Cloud to meet its growing need for computing capacity, sources told Reuters, marking an unexpected collaboration between two fierce rivals in the artificial intelligence space. The deal, finalised in May after months of discussions, adds Alphabet’s cloud services to OpenAI’s infrastructure stack, previously dominated by Microsoft’s Azure.

The move signals a shift in how major AI players navigate the high-stakes battle for leadership in generative AI. Despite being competitors, OpenAI’s ChatGPT poses a serious challenge to Google Search, the two firms have found common ground in the pressing need for vast and scalable computing power.

“We are being super paranoid about safety, so the date could shift,” Elon Musk, CEO of Tesla, recently posted on X when asked about his own AI project’s timeline, a sentiment that echoes across the industry as companies face mounting computational demands.

For Google, the partnership is a notable win. Its cloud business, which contributed $43 billion to Alphabet’s 2024 revenue, is increasingly relying on AI startups and enterprises as growth drivers. By onboarding OpenAI, Google adds one of the most influential AI firms in the world to a customer roster that already includes Apple, Anthropic, and Safe Superintelligence, both of which were founded by former OpenAI leaders.

Scotiabank analysts called the development “somewhat surprising”, noting in a client note that the partnership highlights Google Cloud’s growth potential while also drawing attention to the competitive tension posed by ChatGPT. “The deal underscores the fact that the two are willing to overlook heavy competition between them to meet the massive computing demands,” the analysts said. “Ultimately, we view this as a big win for Google’s cloud unit, but there are continued worries that ChatGPT is becoming an incrementally larger threat to Google’s search dominance.”

OpenAI’s need for compute has ballooned since the viral rise of ChatGPT in late 2022. The company, now on a $10 billion annual revenue run rate, is expanding aggressively with infrastructure initiatives such as the $500 billion Stargate project (in partnership with SoftBank and Oracle) and large-scale deals with CoreWeave. OpenAI is also working on developing its own custom AI chips, with plans to finalise designs later this year.

Until earlier this year, Microsoft served as OpenAI’s exclusive infrastructure partner. But both firms are now in talks to revise the terms of their multibillion-dollar agreement, including Microsoft’s equity stake in the company. Sources say OpenAI’s prior lock-in with Microsoft initially prevented a deal with Google from being signed.

Meanwhile, Google has begun commercialising its own AI hardware, including its tensor processing units (TPUs), which had previously been reserved for internal use. This expansion has allowed it to offer competitive, AI-focused cloud solutions to a broader range of customers. However, Google CFO Anat Ashkenazi acknowledged in April that the company still lacks sufficient capacity to fully meet demand.

Despite the growing user base of ChatGPT and speculation that it could eat into Google’s search market share, Alphabet CEO Sundar Pichai has downplayed the competitive threat. Google’s DeepMind continues to push the frontier with its own advanced models, and the company remains committed to integrating generative AI into both consumer and enterprise products.

Still, selling cloud capacity to OpenAI introduces new complexities, especially in balancing internal resource demands with third-party contracts. As Alphabet’s AI-related capital expenditures are expected to reach $75 billion this year, the pressure is on to deliver returns without compromising strategic autonomy.