Proofpoint announced that it has acquired Acuvity, a Sunnyvale-based AI security and governance startup founded in 2023. Financial terms were not disclosed, though Acuvity’s engineering team is joining Proofpoint as part of the deal. The acquisition is the company’s most pointed move yet into AI-native security, and it signals a broader rethink of where the enterprise security perimeter actually sits in the age of autonomous agents.
The Problem Proofpoint Is Trying to Solve
For most of the past decade, Proofpoint has operated on the premise that the most dangerous thing in any organization’s security posture is a human being: a distracted employee who clicks a phishing link, a disgruntled insider who exfiltrates data, a credentials holder who gets compromised. That framing built a company. It no longer captures the full threat.
AI agents are now active participants in enterprise workflows. They access databases, execute multi-step tasks, connect to external services through APIs and emerging protocols, and make decisions at a speed that has outrun the monitoring tools organizations deployed to watch what humans do. When an autonomous agent queries a cloud drive, calls an external API, or passes sensitive context to a third-party model, the security question is no longer who did this, but whether any system in the organization can even see it happening.
That is precisely the gap Acuvity was built to close. Its platform delivers runtime inspection and enforcement across applications, agents, and Model Context Protocol servers, giving organizations visibility and control across the full range of AI usage in the enterprise, from endpoints and web browsers through to locally installed AI tools like Ollama.
What Acuvity’s RYNO Platform Actually Does
Acuvity launched its flagship RYNO platform out of stealth in June 2025, led by a founding team of cybersecurity veterans. CEO Satyam Sinha previously co-founded Aporeto, a microsegmentation company acquired by Palo Alto Networks, and the rest of the team carried similar pedigrees from senior engineering roles at major security firms.
RYNO is built around four core capabilities. Full Spectrum Visibility provides real-time observability into every AI interaction across an organization’s employees, applications, and agents, surfacing shadow AI use and identifying where sensitive data is entering or leaving AI systems. The Adaptive Risk Engine continuously analyzes AI activity to detect prompt injection attempts, data leakage, unauthorized tool use, and behavioral anomalies as they occur. Context IQ, the platform’s intelligence layer, combines user intent, data classification, application type, and access patterns to understand not just what happened but why it matters. And the Dynamic Policy Engine enforces those assessments in real time, applying guardrails without requiring security teams to write and maintain static rules for every possible scenario.
The sixth capability in RYNO’s feature set is where the platform makes its most forward-looking bet: dedicated security for the Model Context Protocol. MCP is an emerging standard that connects AI assistants to external data sources, enterprise tools, and services, and it is rapidly becoming the connective tissue of agentic AI infrastructure. Acuvity secures MCP by hardening servers with least-privilege execution, immutable runtimes, and continuous vulnerability scanning, while its Minibridge component adds TLS, authentication, and threat detection to close protocol-level gaps. Most enterprise security stacks have no visibility into MCP traffic at all, which makes this a meaningful differentiation rather than a marketing claim.
Why This Matters for Proofpoint’s Platform Strategy
Proofpoint’s Chief Strategy Officer Ryan Kalember described the shift to CyberScoop as a pivot from stopping prompt injection to figuring out what the AI is even doing, which is a candid acknowledgment that the threat model has changed faster than most security vendors have adapted.
The acquisition connects logically to Proofpoint’s May 2025 deal for Hornetsecurity, a Germany-based Microsoft 365 security provider valued at over $1 billion, which expanded Proofpoint’s reach significantly into the mid-market. Kalember has indicated he sees Acuvity extending that reach further, giving smaller organizations that use Hornetsecurity’s infrastructure a credible path into AI security without requiring enterprise-level resources to deploy it.
But the more interesting strategic argument is about what Proofpoint is uniquely positioned to do that pure-play network and data security vendors are not. While network security vendors focus on securing the connection and data security posture management vendors focus on data at rest, Proofpoint is betting that its historical position in human workflows gives it a distinct advantage in securing the payload by understanding not just that an agent is moving data, but why. That intent-aware model is considerably harder to build than it sounds: traditional data loss prevention relies on static rules and pattern matching that increasingly fails against AI systems that can rephrase sensitive content into forms those rules do not recognize.
The Competitive Pressure Is Real
Proofpoint is not making this move in a vacuum. Microsoft, Cisco, Palo Alto Networks, Netskope, Zscaler, and a growing cluster of cloud and data security vendors are all expanding into AI governance territory from their own positions. The window to establish a credible platform story around agentic AI security before that space becomes crowded is narrowing.
What the Acuvity acquisition gives Proofpoint is a purpose-built technical foundation rather than an AI security layer retrofitted onto tools designed for a different era. Role-based access control cannot distinguish between an agent summarizing a document and an agent scanning a cloud drive for credentials, and that limitation runs through most of the existing tooling organizations currently rely on. RYNO was built from the ground up to operate in environments where the actor is not a human being and the behavior is not deterministic, which is a fundamentally different design problem than the one Proofpoint’s legacy platform was built to solve.
The deal does not make Proofpoint’s agentic AI security story complete. What it does is give the company something most of its competitors are still trying to build from scratch: a platform that was designed for autonomous agents from day one, combined with an established enterprise customer base and the distribution to put it in front of them quickly.
