PwC and Google Cloud announced a $400 million agreement to augment the security operations center (SOC) with AI.

The three-year deal expands on the current collaboration between the big four player and Google Cloud Security by utilizing the hyperscaler’s AI-driven security platforms and threat intelligence alongside PwC’s security transformation, risk, and managed services provisions.

Under the new agreement, PwC will extend its agentic, semi-autonomous SOC capabilities to automate investigations and improve visibility across complex environments.

Commenting on the deal, Mauricio Sanchez, senior director at Dell’Oro Group, said the partnership highlights how AI security is increasingly being treated as a shared responsibility across cloud platforms, security tools, and services, rather than a standalone product category.

“Google brings depth in cloud infrastructure, AI models, and native security controls, while PwC adds governance, risk, and operational expertise that many enterprises lack as they move AI from experimentation into production,” Sanchez said. “The focus on secure-by-design AI and model governance aligns with what we are hearing from enterprises that are now prioritizing scale and risk management over pilots.”

PwC and Google’s collaboration follows on from recent AI-centric deals in the cybersecurity space, such as Cato Networks buying Aim Security to add AI security to its secure access service edge (SASE) services, and Palo Alto Networks’ mega-acquisition of CyberArk to filter out human and agentic identities. In a similar deal to the PwC one, Palo Alto expanded its own strategic partnership with Google Cloud in December.

Out of PwC's rivals in the consultancy space, Accenture has deepened its AI security focus by buying U.K. player Faculty for in-baked secure AI solutions this month, while Deloitte and Amazon Web Services partnered with an agentic data trust platform from Canada’s Ataccama.

“This reinforces that AI security is becoming embedded within cloud and security architectures rather than emerging as a discrete point solution,” added Sanchez. “Partnerships like this are less about near-term revenue and more about shaping reference architectures and buyer expectations around how AI workloads should be secured and governed, particularly as regulatory scrutiny continues to increase.”