Palo Alto and Google Cloud seal nearly $10B AI security pact

This is one of the largest publicly reported AI‑security tie‑ups so far, and it underscores how fast defensive tooling is having to evolve to keep pace with model deployment. By standardising on Google Cloud as both infrastructure and AI stack, Palo Alto is effectively betting that hyperscalers will become the default substrate for securing agentic AI as much as for running it. In return, Google gets a powerful validation that its Vertex/Gemini ecosystem can host the security companies that protect the rest of the enterprise world.

For the AGI race, the deal is a reminder that whoever wins at models must also win at securing them. The more organisations push sensitive workloads onto AI systems, the more catastrophic a successful compromise becomes. Embedding Palo Alto’s Prisma AIRS platform deep into Google Cloud’s pipelines—code to cloud, AI posture management, runtime monitoring—helps normalise the idea that every AI deployment needs a first‑class security posture, not bolt‑on firewalls. It also tightens Google’s competitive grip versus AWS and Azure in security‑sensitive AI workloads, which could influence where future frontier‑model training and inference clusters end up running.