Google Cloud claims its AI chip business is twice as large as a rival's. What builders need to know.

Google Cloud CEO Thomas Kurian made a bold claim at a Goldman Sachs conference: the company's AI accelerator chip business is twice as large as a major cloud computing rival's. For builders, that's a statement worth watching, but the more concrete number is the cloud revenue growth that backs it. Google Cloud reported $24.77 billion in Q2 revenue, up 82% year over year, driven by AI workloads and beating analyst estimates of $22.46 billion.

What Kurian actually said

At the Goldman Sachs conference, Kurian stated that Google's emerging business of selling AI accelerator chips is more than twice the size of a competitor. The exact rival was not named in the report, but the implication is that Google is positioning its custom TPU (Tensor Processing Unit) business as a major force. Google has been designing its own chips for years, and now sells access to them through cloud services. The company also provides TPUs to OpenAI and hosts Anthropic's Claude on Vertex AI, which gives it a unique position as both a chip supplier and a model competitor.

Why the cloud revenue growth matters more

The $24.77 billion cloud revenue figure is the strongest signal that Google's AI infrastructure investment is paying off. The 82% year-over-year growth, beating estimates by over $2 billion, suggests that enterprises are moving significant AI workloads to Google Cloud. For builders, that means more capacity, competitive pricing, and continued investment in the platform. It also means Google is less dependent on third-party chip suppliers, which could lead to better cost structures for AI inference and training.

What builders should watch

If Google's chip business is indeed that large, it could put downward pressure on cloud AI pricing across the industry. Google has a history of passing efficiency gains to customers. But the claim remains unverified, and the chip business size is not directly comparable to cloud revenue. Builders should monitor Google's next earnings call for more granular hardware disclosures. The more relevant takeaway is that Google Cloud is becoming a serious contender for AI workloads, especially for teams that want to use Gemini or access Anthropic through Vertex.

The caveat

The claim about chip business size comes from a single executive statement at a conference, not from audited financials. Kurian's comments are part of a competitive narrative, and without independent data, it's hard to gauge the actual scale. The linked source notes that the article is based on Kurian's claim and media coverage, not independently verified figures. For now, treat it as a directional signal, not a hard fact.