Google Cloud Bets on AI Startups as Nvidia–OpenAI Mega Deal Reshapes Market

Google Cloud expands AI infrastructure to capture next-generation startups

The artificial intelligence infrastructure race is accelerating at breakneck speed. Just days after Nvidia and OpenAI’s record-breaking $100 billion deal, Google Cloud is pursuing a different strategy — one that could define the second wave of AI.

Nvidia–OpenAI Mega Deal Reshapes AI Infrastructure

On Monday, Nvidia and OpenAI revealed a $100 billion partnership involving massive chip purchases and computing power equivalent to more than 5 million U.S. households. This deal highlights the consolidation trend in AI infrastructure, as major players secure exclusive, long-term partnerships.

Microsoft has already poured nearly $14 billion into OpenAI, Amazon invested $8 billion in Anthropic, and Oracle recently landed a $30 billion cloud deal with OpenAI before signing a staggering $300 billion five-year commitment starting in 2027. Even Meta, despite building its own infrastructure, struck a $10 billion agreement with Google Cloud while planning $600 billion in U.S. spending through 2028.

These massive commitments raise the question: Is Google Cloud being left out?

Google Cloud’s Startup-Centric Strategy

Rather than competing for today’s AI giants, Google Cloud is betting on AI startups and future unicorns. According to COO Francis deSouza, Google is focused on becoming the primary cloud provider for next-generation AI companies before they scale too big.

Key figures from deSouza:

• 9 of the top 10 AI labs use Google Cloud infrastructure.
• Nearly all generative AI unicorns run on Google Cloud.
• 60% of global generative AI startups rely on Google as their cloud provider.
• $58 billion in new revenue commitments secured over the next two years — more than double its current annual run rate.

Google offers startups $350,000 in credits, technical support, enterprise-grade infrastructure, and a “no-compromise AI stack” that spans chips, models, and applications.

Google’s TPU chips power advanced generative AI infrastructure worldwide.

Google’s Expanding AI Infrastructure

Beyond customer acquisition, Google is scaling its custom TPU chip business. Recent reports reveal deals to place TPUs in other cloud providers’ data centers, including a $3.2 billion partnership with London-based Fluidstack to back a New York facility.

Google Cloud also provides TPUs to OpenAI and hosts Anthropic’s Claude model on Vertex AI, even while competing directly with its Gemini AI models. DeSouza frames this as Google’s “open ethos” — offering technology to rivals while fostering competition.

This philosophy aligns with Google’s open-source legacy, from Kubernetes to the transformer architecture behind modern AI. More recently, Google published its Agent-to-Agent (A2A) protocol to standardize inter-agent communication.

Regulatory Pressure Meets AI Opportunity

Google’s aggressive push into AI comes amid scrutiny. A federal court ruling earlier this month in the search monopoly case attempted to limit Google’s market dominance without hindering its AI ambitions. While Google avoided harsher penalties, regulators remain concerned about the company leveraging search data to dominate AI.

Still, deSouza emphasizes the positive potential of AI. He cites opportunities in Alzheimer’s, Parkinson’s, climate science, and drug discovery, arguing Google Cloud can power breakthroughs in healthcare and sustainability.

Google Cloud Bets on Tomorrow's Giants

In a market where startups can scale into multi-billion-dollar companies almost overnight, capturing tomorrow’s AI leaders may prove more valuable than chasing today’s giants. By positioning itself as an open, supportive partner for AI startups, Google Cloud strengthens its competitive edge while signaling to regulators that it fosters innovation rather than stifles it.

As deSouza puts it: “AI is resetting the cloud market, and Google Cloud is leading the way — especially with startups.”

---