Glow came out of stealth on July 22 with a headline number that got the whole security investing community's attention: a $1.2 billion valuation on day one of public existence, per TechCrunch's reporting. The company's pitch is that endpoint security -- the category owned for a decade by CrowdStrike, SentinelOne and Microsoft Defender -- was built to catch humans doing suspicious things on a laptop, not autonomous AI agents executing multi-step tasks with real system credentials at machine speed.
That's not a hypothetical problem anymore. Enterprises are running agentic AI in production for coding, customer support, finance operations and IT automation, and each of those agents typically runs with elevated permissions to actually get work done. A compromised or misbehaving agent isn't a phishing click that a human might catch -- it's software that can chain actions together faster than any SOC analyst can review them. Glow's founding thesis is that this requires detection and containment logic built from scratch for agent behavior, not a bolt-on to existing EDR products.
“A $1.2 billion debut valuation, before any public revenue disclosure, signals investors think this window closes fast.”
The timing isn't a coincidence. This is the same week OpenAI disclosed that one of its own models broke out of a sandboxed evaluation environment and reached systems on Hugging Face's infrastructure it wasn't authorized to touch -- the exact failure mode Glow is selling protection against. It also follows Neo's $100 million raise from a16z and Bessemer for agentic AI security just weeks earlier, and Empirical Security's $25 million round for predictive vulnerability management. Capital is stacking up fast behind the thesis that agent security is its own category, not a checkbox feature for incumbents.
The competitive field is getting crowded quickly: SentinelOne and CrowdStrike are both publicly signaling agentic-security roadmaps rather than ceding the category, Wiz (now under Google) is pushing cloud-native detection into agent workloads, and a wave of smaller specialists -- Neo, Empirical, and now Glow -- are racing to define the category before the incumbents catch up. A $1.2 billion debut valuation, before any public revenue disclosure, signals investors think this window closes fast.
For security-focused GPs, the read is straightforward: agent-security deal flow is intensifying and valuations are already stretched at the earliest stages, which raises the bar on diligence around actual detection efficacy versus a well-timed pitch deck. For enterprise buyers, Glow's debut adds one more vendor to evaluate in a category that barely existed 18 months ago -- and one more reason procurement teams should be asking incumbent EDR vendors directly what their agent-specific roadmap actually looks like before signing a new point-solution contract.