Glow's $180M says the laptop is the new AI perimeter

Investors just paid a $1.2 billion valuation for a company with no public product, betting on one idea: the employee laptop became the most dangerous machine in your company the day AI agents started running on it. Glow emerged from stealth July 22 with a $180 million Series A to rebuild endpoint security for that world. You almost certainly can't buy it yet. You can act on the thesis today.

What actually happened

Per TechCrunch and Help Net Security, Glow raised $180M in an all-equity Series A at a $1.2B valuation, led by Sequoia Capital, Cyberstarts, Greenoaks, and Redpoint, with Index, Lux, Swish, and Operator Collective participating.

Founded in 2025 and headquartered in Palo Alto, the team is unusually senior: CEO Roi Tiger was a VP of Engineering at Meta, and co-founders came out of Snowflake's security strategy org and Claroty's R&D. COO Emily Heath was CISO at United Airlines and DocuSign.

The product uses agents — running on Anthropic and Google Gemini models via Amazon Bedrock — to map what's actually running on employee devices and enforce policy. The concrete examples TechCrunch cites are the tell: blocking malicious npm packages at install, catching AI agents pulling risky software, and finding machines where endpoint detection is missing or switched off. Deployments reportedly span tens of thousands of devices across healthcare, retail, and financial services. Tiger's framing: for a decade everything moved to cloud and SaaS, and now "AI lands on the endpoint in a way we've never seen."

Why endpoint security matters for your business now

Because the threat model here is not exotic. It's a developer laptop running a coding agent with terminal access, an MCP server someone installed from a README, and an npm install that executes a postinstall script. We wrote about npm 12 blocking install scripts and the Cursor RCE for exactly this reason. A $180M startup is now the market's answer to a problem your five-person team already has.

You don't need a unicorn's platform to close most of the gap. Write down which AI tools are approved and which machines run them — you cannot secure an inventory you don't have. Turn off install scripts by default in CI and on dev machines. Give agents scoped, short-lived credentials instead of the developer's personal token, so a compromised session can't reach production. Require approval mode for anything that shells out. And confirm your existing EDR is actually running on every device, because "missing or disabled" is the finding Glow is selling against.

That's a week of work, not a $1.2B platform.

Sources: TechCrunch, Help Net Security.