The cybersecurity veterans didn't need much of a pitch.
Pavel Gurvich and three of his Guardicore colleagues—the ones who'd helped build a company that Akamai bought for roughly $600 million four years ago—walked into rooms at Greylock Partners, Battery Ventures, and Lux Capital with a proposal: Give us money to build artificial intelligence that attacks software the way nation-states do. Continuously. Autonomously. At scale.
By November 11, those conversations had turned into $75 million, one of the fattest seed checks the cybersecurity world has seen. Tenzai, the Tel Aviv startup emerging from stealth that day, now carries a valuation hovering around $330 million—remarkable for a company that won't actually start selling its product until early next year.
Five months earlier, the company had already quietly banked $32 million, according to Israeli outlet Calcalist. That initial haul, combined with this latest infusion, speaks to something venture capitalists rarely admit out loud: pedigree still opens wallets faster than traction.
The Guardicore Effect
Four of Tenzai's five founders earned their stripes at Guardicore, a network security company that became one of Israel's notable exits when Akamai snapped it up in 2021. Gurvich, now CEO, led that effort alongside Ariel Zeitlin (Tenzai's CTO), Ofri Ziv (VP of Research), and Itamar Tal (VP of Engineering). The fifth wheel is Aner Mazur, who spent time as chief product officer at Snyk before a stint as entrepreneur-in-residence at Accel in 2024—a role that often functions as a well-paid holding pattern for founders plotting their next move.
Together, they're now running an 18-person operation. Not exactly a sprawling empire, but in the compressed timeline of AI-era startups, perhaps more than the founders expected at this stage.
What Tenzai Actually Does
Strip away the jargon and Tenzai is building software that hacks other software—constantly, methodically, the way elite security teams (or hostile governments) might probe for weaknesses. The company calls it "agentic AI penetration testing," which translates to autonomous agents powered by frontier models from Anthropic and OpenAI. These systems don't just scan for known vulnerabilities. They exploit them, document how it's done, and theoretically help companies patch the gaps before someone malicious does the same.
"Pentesting hasn't changed much," Gurvich said in the announcement, leaning into a familiar gripe among security professionals. Traditional penetration testing—where human experts manually probe systems for flaws—tends to happen episodically, maybe once or twice a year if a company is diligent. Tenzai's pitch is making that process continuous, software-driven, always-on.
"AI allows us to bring elite, nation-grade offensive capabilities... always-on offensive capabilities," Gurvich added, invoking the specter of state-sponsored hacking with language calibrated to unsettle corporate security chiefs.
The market they're chasing is substantial—roughly $8 billion globally for penetration testing services. But it's also crowded and fragmented, with boutique consultancies, established vendors, and now a wave of AI-focused startups all vying for the same budgets.
Early Traction, Light on Details
Tenzai says it's already deployed early versions of the platform with unnamed enterprises in financial services, healthcare, and technology. The company remains in private beta, a designation that carries less weight than it once did now that "beta" has become a perpetual state for many software products. Full commercial sales are slated to begin in early 2026, which gives the team roughly a quarter to refine what they're building and figure out how to price it.
That timeline also suggests investors are betting on potential rather than proof. Asheem Chandna, the Greylock partner who led the round, framed the investment around a familiar anxiety: "AI generated code is creating new cybersecurity risks." It's a thesis that resonates in an era when developers are leaning heavily on tools like GitHub Copilot and ChatGPT to write software faster—often without fully understanding what the AI is producing.
Whether Tenzai's approach actually solves that problem, or simply creates a new category of tool that security teams will need to learn, remains an open question.
What Comes Next
The $75 million will fund the usual startup expansion playbook: hire more AI researchers and security engineers, beef up the offensive capabilities of the platform, and build out sales and marketing operations in North America and Europe. Tenzai joins a crowded field of Israeli cybersecurity firms trying to crack U.S. enterprise budgets, though the Guardicore pedigree likely smooths some of those introductions.
Forbes first reported the $330 million valuation in November, a figure that feels aggressive given how little revenue—if any—the company has generated so far. Then again, cybersecurity has always been a sector where fear drives faster decisions than logic. If Tenzai can convince enough CISOs that AI-generated code is a ticking time bomb, maybe the valuation starts to make sense.
Or maybe it's just another data point in the frothy fundraising environment that surrounds anything with "AI" and "autonomous" in the same sentence. The product, eventually, will have to speak for itself.
