Founded last year by veterans of Meta and Snowflake, the startup arrives as an instant unicorn with a pointed thesis: the AI agents now running on employee laptops have broken the assumptions that endpoint security was built on.

A cybersecurity startup called Glow came out of stealth on Wednesday with one of the largest Series A rounds of the year, announcing 180 million dollars in all-equity funding at a 1.2 billion dollar valuation. The company, headquartered in Palo Alto with most of its engineering in Israel, was founded in 2025 and has spent its short life building a platform that watches and controls what runs on employee devices, with a particular focus on the autonomous AI agents enterprises are now installing at scale.

The pitch lands at a moment when AI sits on both sides of the security equation, powering the attackers and multiplying what defenders have to protect.

An Instant Unicorn in Security's Hottest Corner

The round was led by Sequoia Capital, Cyberstarts, Greenoaks, and Redpoint Ventures, with participation from Index Ventures, Swish Ventures, Lux Capital, Operator Collective, and Holly Ventures. Glow says the money will fund an expanded go-to-market push in the United States and the growth of Glow Labs, its internal security research division.

Crossing the billion-dollar line before publishing a single revenue figure would once have raised eyebrows. In 2026 it has become a pattern in AI security. Tenex passed a billion-dollar valuation in April on the back of a 250 million dollar round for AI-managed detection and response, and XBOW, started by GitHub Copilot creator Oege de Moor, reached unicorn status for its AI systems that hunt software vulnerabilities. Investors are treating the intersection of AI and security as land worth grabbing before the map is drawn.

Glow says it already has paying customers across industries including healthcare, retail, financial services, and others, with typical deployments covering tens of thousands of devices inside global organizations.

It has declined to name any of those customers or say how many there are.

The Résumés Behind the Valuation

The valuation is, in large part, a bet on the founding team. Chief executive Roi Tiger spent nine years at Meta and rose to vice president of engineering, having arrived through Facebook's 2013 acquisition of Onavo, the mobile analytics startup he co-founded. Onavo later became a cautionary tale, shut down after criticism of its data-collection practices, which gives Tiger an unusual credential for a security founder: firsthand experience of how monitoring tools can cross lines.

His co-founders bring depth from different corners of the field. Chief technology officer Omer Singer previously led cybersecurity strategy at Snowflake, where he championed the security data lake concept, and served earlier in Israel's Unit 8200 intelligence corps. Ophir Arie, who runs research and development, held the same post at industrial cybersecurity firm Claroty. A fourth co-founder, Arnon Joseph, is a longtime Meta product and engineering leader.

Rounding out the executive bench is chief operating officer Emily Heath, formerly the chief information security officer of United Airlines and Docusign. Heath sat on the board of Wiz through its 32 billion dollar acquisition by Google and was previously a partner at Cyberstarts, the same firm now co-leading Glow's round.

What the Platform Actually Does

Glow describes its product as a control layer over everything running on an employee device, from ordinary applications to developer tools and autonomous AI agents. Specialized AI agents of Glow's own continuously map the enterprise environment and score risk as conditions change. A policy engine then acts on that picture, deciding which software is allowed in and which gets removed.

Under the hood, the platform runs on models from Anthropic and Google's Gemini, accessed through Amazon Bedrock, with Glow's own software feeding those models enterprise context and tightening their reliability for security work.

Tiger says the system has already earned its keep in early deployments. It has blocked malicious npm packages, the third-party code components developers fold into applications, and caught AI agents in the act of trying to install them. It has also flagged employee devices where endpoint detection and response tools were missing entirely or running in a weakened state, a quiet failure mode that attackers actively hunt for.

"Suddenly, AI lands on the endpoint in a way we've never seen," Tiger said in an interview with TechCrunch, contrasting the moment with the previous decade's migration of everything to cloud and SaaS.

Why Now: AI on Both Sides of the Device

The offense half of Glow's thesis is familiar by now. Attackers are using generative AI to automate phishing at scale and to speed up malware development, compressing work that once took specialist teams into something closer to a subscription.

The defense half is newer. Enterprises are placing AI coding assistants and autonomous agents directly on employee machines, and those agents do things no ordinary application does: they fetch code from public repositories, execute commands, and act without a human clicking approve. Each one is a productivity gain and a potential unsupervised actor on the corporate network at the same time.

Anxiety about capability has sharpened the debate this year. Anthropic disclosed in April that it was holding back a preview of its Mythos model over concerns about how good it had become at breaking into software, and the model's eventual unveiling, with what the company described as advanced ability to find and exploit vulnerabilities, intensified discussion of AI-assisted intrusion across the industry. When a frontier lab treats its own model's hacking skill as a risk to be managed, enterprise security teams take note.

A Crowded Field Ruled by Giants

Glow is walking into one of the most contested markets in software. Endpoint security is dominated by CrowdStrike, Microsoft, SentinelOne, and Palo Alto Networks, all of them established, well-funded, and shipping AI features of their own.

Tiger's differentiation argument is about posture. Existing endpoint detection and response products, he says, concentrate on spotting threats after they surface, while Glow is built to stop risky software and unauthorized agents from getting onto machines in the first place. In comments to Calcalist, he framed the ambition wider still, describing a single platform aimed at the entire endpoint market rather than an AI-specific add-on, replacing the patchwork of point products enterprises currently stack on every laptop.

The company remains small for the fight it has picked, employing close to 100 people, roughly 70 percent of them in Israel and the rest in the United States.

The unresolved question hanging over the launch is whether AI-native endpoint security becomes a category of its own or gets absorbed as a feature of the platforms enterprises already pay for. Glow enters that contest with no published revenue and no named customers. What it has instead is a founding team the industry already trusts and a thesis that the next wave of AI-driven breaches will either confirm or embarrass.