Market landscapes / AI Safety, Security & Governance
AI Runtime Security & Guardrails companies
Guardrails that run at the moment a model is being used: blocking prompt injection, stopping data leaving, and enforcing policy on what a model may say and do. 12 of the 74 have been acquired, more than in any other sub-sector on these maps. Runtime security keeps becoming part of a bigger security product rather than staying a purchase of its own.
Updated 12 October 2026
74 companies
12 acquired
38 hiring now
The rest of the AI Safety, Security & Governance landscape
The full map, including the sub-sectors too small to list separately.
The whole AI Safety, Security & Governance mapAI Governance, Risk & Compliance (GRC) 68AI Red Teaming & Security Testing 39Agent Identity & Access Control 32Shadow AI Discovery & Workforce AI Governance 29AI Security Posture & Threat Detection 24AI Data Privacy & Confidential Computing 21
I am Ben Jackson. I map this market and I hire the commercial teams inside it: sales, marketing, product and delivery. If you are building a team in AI Runtime Security & Guardrails, here is how I work. The whole market in numbers.