it isn’t your imagination: google cloud is flooding the zone

A $100 Billion AI Partnership and Google’s Alternative Strategy

The recently announced $100 billion collaboration between Nvidia and OpenAI signifies, at present, the newest substantial agreement reshaping the AI infrastructure sector and strengthening the ties between two leading AI entities.

Concurrently, Google Cloud is pursuing a distinctly different approach. While major industry players are solidifying increasingly close partnerships, Google Cloud is determined to secure the next wave of AI companies before they reach a size that makes them difficult to engage with.

Google Cloud’s Focus on the Next Generation

Francis deSouza, the COO of Google Cloud, has observed the AI revolution from various perspectives. Previously as the CEO of genomics leader Illumina, he witnessed machine learning’s impact on drug development. As a co-founder of Synth Labs, a two-year-old AI alignment startup, he has confronted the safety concerns associated with increasingly sophisticated models. Now, in his role at Google Cloud since January, he is leading a significant investment in AI’s subsequent phase.

DeSouza frequently illustrates this with data. He highlighted to our team that nine of the top ten AI labs utilize Google’s infrastructure. He also stated that almost all generative AI unicorns operate on Google Cloud, with 60% of all GenAI startups globally selecting Google as their cloud provider. Furthermore, the company has secured $58 billion in new revenue commitments over the next two years, exceeding its current annual revenue.

When questioned about the proportion of Google Cloud’s revenue derived from AI companies, he emphasized that “AI is fundamentally changing the cloud market, and Google Cloud is at the forefront, particularly with startups.”

Consolidation in AI Infrastructure

The Nvidia-OpenAI agreement exemplifies the extensive consolidation occurring within AI infrastructure. Microsoft’s initial $1 billion investment in OpenAI has expanded to nearly $14 billion. Amazon responded with $8 billion in investments in Anthropic, securing customized hardware that optimizes AI training for Amazon’s infrastructure. Oracle has also emerged as a key player, securing a $30 billion cloud deal with OpenAI and a subsequent $300 billion five-year commitment starting in 2027.

Even Meta, despite developing its own infrastructure, entered into a $10 billion agreement with Google Cloud while simultaneously planning $600 billion in U.S. infrastructure investments through 2028. The Trump administration’s $500 billion “Stargate” project, involving SoftBank, OpenAI, and Oracle, adds another dimension to these interconnected partnerships.

Google’s Strategy: Nurturing Emerging Companies

These substantial deals could appear as a challenge for Google, considering the partnerships that OpenAI and Nvidia are establishing elsewhere. However, it seems Google is not remaining passive. Instead, Google Cloud is engaging with smaller companies like Lovable and Windsurf – the “next generation of companies,” as deSouza describes them – without significant upfront investments.

This strategy goes beyond simply acquiring customers. Google provides AI startups with $350,000 in cloud credits, access to its technical teams, and support for market entry through its marketplace. Google Cloud also offers a comprehensive AI stack – encompassing chips, models, and applications – with an “open ethos” that provides customers with flexibility at every level.

This approach is driven by both opportunity and necessity. In a market where companies can rapidly grow “from a startup to a multibillion-dollar company,” as deSouza points out, securing future unicorns before they mature may prove more valuable than competing for today’s established leaders.

“Companies appreciate access to our AI stack and our teams, providing insights into our technology roadmap,” deSouza explained. “They also value access to Google’s enterprise-grade infrastructure.”

Expanding Google’s AI Chip Business

Google’s infrastructure strategy has become even more ambitious, with reports revealing its efforts to expand its custom AI chip business. According to The Information, Google has reached agreements to deploy its tensor processing units (TPUs) in other cloud providers’ data centers, including a deal with London-based Fluidstack, backed by up to $3.2 billion for a New York facility.

Balancing direct competition with AI companies while simultaneously providing them with infrastructure requires careful management. Google Cloud supplies TPU chips to OpenAI and hosts Anthropic’s Claude model via its Vertex AI platform, even as its own Gemini models compete directly with both. (Alphabet, Google Cloud’s parent company, holds a 14% stake in Anthropic, according to New York Times court documents; when asked about Google’s financial ties to Anthropic, deSouza described the relationship as a “multi-layered partnership,” then directed attention to Google Cloud’s model marketplace.)

A Commitment to Openness

If Google aims to be a neutral party while advancing its own interests, it has considerable experience. This approach is rooted in Google’s open-source contributions, from Kubernetes to the foundational “Attention Is All You Need” paper that enabled the transformer architecture powering most modern AI. Recently, Google released an open-source protocol called Agent2Agent (A2A) for interagent communication, demonstrating its ongoing commitment to openness even in competitive areas.

“We have consistently chosen to be open at every layer of the stack, recognizing that companies may leverage our technology to build competing solutions,” deSouza acknowledged. “This has been happening for decades, and we accept it.”

Navigating Regulatory Scrutiny

Google Cloud’s focus on startups occurs at a critical juncture. This month, Judge Amit Mehta issued a ruling in the government’s five-year-old search monopoly case, aiming to curb Google’s dominance without hindering its AI ambitions.

While Google avoided the Justice Department’s most severe proposed penalties, including the divestment of Chrome, the ruling highlighted regulatory concerns about Google leveraging its search monopoly to dominate AI. Critics fear that Google’s extensive search data provides an unfair advantage in AI development and that the company might employ similar monopolistic tactics that established its search dominance.

DeSouza emphasized positive potential outcomes. “We have an opportunity to gain a fundamental understanding of major diseases like Alzheimer’s and Parkinson’s, and to advance climate technologies,” he stated, outlining a vision where Google Cloud supports research in these areas. “We are dedicated to pioneering the technologies that will enable this work.”

Critics may remain skeptical. By positioning itself as an open platform that empowers the next generation of AI companies, Google Cloud may demonstrate to regulators that it fosters competition, while simultaneously building relationships with startups that could strengthen Google’s position if regulatory pressure increases.

Listen to our complete conversation with deSouza on this week’s StrictlyVC Download podcast; a new episode is released every Tuesday. Please note a correction as of 9/26: We incorrectly reported deSouza as calling Google Cloud the “primary” computing partner of startups Windsurf and Lovable; we apologize for the error.