The list. Glow made the 2026–2027 Fortune Cyber 60 in the early-stage group. Fortune’s profile lists it as founded in 2025, headquartered in Palo Alto, with 128 employees and CEO Roi Tiger. Fortune reports $180 million in funding. Glow came out of stealth on July 22 saying it raised that $180 million at a $1.2 billion valuation, led by Sequoia, Cyberstarts, Greenoaks and Redpoint. Lightspeed, which helps run the Cyber 60, is not among the investors Glow named.
Who built it. Tiger spent nine years at Meta, most recently as a VP of engineering, after Meta bought Onavo, the company he founded. CTO Omer Singer was Snowflake’s head of cybersecurity strategy. Ophir Arie ran R&D at Claroty. The COO is Emily Heath, former CISO at United Airlines and DocuSign. That is a team built to sell to CISOs, not just to impress engineers.
The problem, in plain words. The laptop is where AI enters the company. Employees install desktop assistants, coding agents and command-line tools faster than anyone can review them. Coding agents pull in open-source packages on their own. Glow’s own site claims the average organization runs more than 12,000 unique pieces of software and that 67% of it is invisible to existing tools. Those are company figures, not audits. Directionally, every CISO knows the feeling. EDR, endpoint detection and response, is very good at spotting malicious behavior once something runs. It was never designed to decide whether that thing should have been there in the first place.
The unique part: prevention that doesn’t break the business. Allowlisting, meaning only approved software can run, has always been the most secure endpoint control. It has also always failed, because humans can’t review thousands of apps and the business revolts when work gets blocked. Glow’s bet is that AI agents can do the review. Its agents build a live inventory of every device and every layer of software, research and reconcile what they find, judge the risk against the customer’s own policies, then remove or block software automatically within limits the security team sets. Tiger told TechCrunch the platform has already stopped malicious npm packages, the open-source building blocks developers use, from being installed, caught AI agents trying to pull such packages in, and found devices where the EDR agent was missing or running in a degraded state. That last one is underrated. A control that isn’t running is the most common gap in any fleet.
The proof, and its limits. Glow’s homepage quotes CISOs at Fanatics, CAVA and Antares Capital and the CIO of BMC Software. It claims an average of 28 days to zero critical issues and a 90%-plus reduction in AI supply-chain risk. Those are company claims and testimonials. Glow told TechCrunch it has paying customers in healthcare, retail and financial services, with deployments spanning tens of thousands of devices, but declined to give customer counts or revenue.
Question the valuation. $1.2 billion for a company founded in 2025 with no disclosed revenue is a bet on the team and the category, not on numbers. Even the round story varies. TechCrunch called it a $180 million all-equity Series A. Ctech reported three rounds in about a year: a $20 million seed, a $60 million Series A at a $400 million valuation and a $100 million Series B. Treat the structure as reported, not settled. Tiger pitched Calcalist on a $40 billion endpoint market dominated by CrowdStrike and SentinelOne. That is his figure, and it describes the whole market, not what Glow can win. Glow does not need to replace EDR to justify its price. It needs to become the policy layer that sits above it.
Why it’s winning now. AI moved work back onto the device just as attackers got faster tools. Glow’s launch release says defenders should assume attackers have Mythos-class capabilities, meaning frontier AI that finds and exploits flaws at machine speed. If you accept that, every risky app you tolerated becomes a ticking clock, and cleaning up before the incident beats responding after it. The risk is the platform squeeze. CrowdStrike and Microsoft already have an agent on most corporate laptops, and both can add software-control features. Glow has to be clearly better at the decision, not just first to market.
For practitioners. Run Glow’s inventory against your own asset list for a week. The gap between what you think runs on your fleet and what actually runs is your real attack surface, and the cheapest number you will ever get from a pilot.
Would you let an AI agent automatically remove software from employee laptops, or does every block still need a human sign-off in your shop?