Q&A with Google Cloud CEO Thomas Kurian on building infrastructure for AI agents, balancing internal needs and the demands of customers like Anthropic, and more
An interview with Google Cloud CEO Thomas Kurian about Google's cloud priorities, enterprise agent platform, and Google's integration advantage.
Context & Ripple Effects
Kurian’s earlier interviews framed Google Cloud’s AI push around a competitive reset, efficient training and inference, and a broad model portfolio. By late 2025, the discussion had moved up the stack to Gemini Enterprise and the practical limits of agent hype.
This interview extends that arc from models and enterprise software to the infrastructure needed to run agents, while highlighting the tension between Google’s own AI requirements and those of major cloud customers such as Anthropic.
First-order effects
- Google Cloud’s infrastructure and enterprise-agent priorities become more explicitly coupled: capacity, platform design, and product work must serve both agent workloads and customer demand.
- Anthropic and similarly demanding cloud customers gain added strategic importance because their requirements help shape the infrastructure Google Cloud is building alongside its internal needs.
Second-order effects
- Google Cloud’s ability to offer a credible agent platform will depend on translating its internal integration advantages into customer-facing services, rather than merely operating those capabilities inside Google.
- AWS and Microsoft face continued pressure to compete on the full agent stack—underlying infrastructure, model access, and enterprise deployment—rather than treating compute capacity as a standalone offering.
Third-order effects
- If agent workloads become a durable cloud demand category, AI competition will increasingly center on integrated infrastructure platforms that reconcile first-party AI use with external-customer flexibility.
- The balance between serving internal models and large external AI customers may become a defining constraint on cloud differentiation, with the strongest platforms needing to prove that internal integration does not narrow customer choice.
The trend: This is one data point in AI infrastructure platformization: cloud providers are turning agent support into an integrated contest across capacity, models, enterprise tools, and operational know-how.