Startup Astelia Secures $35M for AI Vulnerability Management

An exposure management startup run by the ex-leader of Israel’s National Red Team raised $35 million to help organizations focus on the most important vulnerabilities.

See Also: AI or Data Governance? Gartner Says You Need Both

The Index Ventures-led Series A funding round will help New York-based Astelia reduce vulnerability noise by identifying the small percentage of vulnerabilities that are truly exploitable within a specific network context, said co-founder and CEO Alon Noy. Astelia currently focuses on prioritization and remediation but plans to expand into vulnerability scanning to become an end-to-end platform.

“Because of artificial intelligence, the amount of vulnerabilities increases dramatically,” Noy told Information Security Media Group. “The amount of the time to exploit shrinks towards zero. Sometimes the same day the CVE is published, an exploit is already over there in the wild. The traditional problem of vulnerability management is not something new. This problem was there always. I think now it’s worse than ever.”

Astelia, founded in 2024, has been led since its inception by Noy, who previously spent more than eight years leading Israel’s National Red Team, the unit responsible for stress testing the country’s critical infrastructure defenses. Company co-founders Nadav Ostrovsky and Roy Rajwan led elite Israeli Red Teams focused on real-world attack paths, adversary behavior and large-scale defensive systems.

Understanding the Exact Conditions Required for Exploitation

Lead Series A investor Index Ventures brings experience in scaling iconic technology companies, Noy said, while lead seed investor Team8 provided early-stage mentorship, customer access through its CISO Village and ideation-stage validation. Noy said Astelia wants to double R&D capacity to deepen network analysis, expand scanning capabilities and strengthen AI-driven product features.

“When such an amazing VC like Index gives you that offer, and you know that that would enable you to boost go-to-market in the U.S. and double your R&D team, this is something you need to think carefully not to take it,” Noy said. “It was a great opportunity, and I think it will enable us continue the journey.”

Attackers now possess capabilities that dramatically compress the time between vulnerability disclosure and exploitation, meaning defenders no longer have a buffer period to assess and remediate issues, Noy said. The longstanding “needle in a haystack” problem becomes existential when attackers are moving at machine speed, and Noy said defensive AI must now match offensive AI.

“I’ve been attacking the good guys as part of my job to find the gaps and then help them protect that,” Noy said. “I wish that in my days on the National Red Team, I could have the same agent AI capabilities that are available today for attackers.”

Astelia’s AI agents deeply analyze each vulnerability’s technical prerequisites to produce structured intelligence about the exact conditions required for exploitation, and the company also simulates real-world network connectivity to determine reachability by mapping network topology and exposure paths. By cross-correlating, Astelia can reduce the actionable vulnerability set to just 1% to 2% of findings.

“The vulnerability analysis lets us understand what are the requirements, the relevant port process, all the things that are locally in the machine. But then without analyzing the network and proving if the full chain exists or not, it doesn’t really matter when you’re actually blind,” Noy said. “When you have both of them, you’re able to have real reachability analysis, especially in such a large environment.”

Why Astelia Wants to Get Into First-Party Scanning

While Astelia currently integrates with third-party scanners such as Tenable, Rapid7, Qualys and CrowdStrike, Noy outlines a clear ambition to expand into first-party scanning and use its contextual awareness to narrow scanning scope and improve signal quality. Because the company already performs deep network modeling and exploitability analysis, it can focus scanning efforts more intelligently.

“The unique network approach that we bring to the table in exposure management provides something else,” Noy said. “We reduce the noise much better than others because of the network analysis. We use agentic remediation to provide multiple remediation plans beyond patching.”

Hybrid environments contain diverse vendors, cloud providers and on-premises systems, and Noy said accurate modeling requires simulation of traffic paths and enforcement rules across heterogeneous infrastructures. Neither vulnerability vendors nor network firewall vendors have mastered this capability since vulnerability vendors lack network depth, and network vendors lack exploitability intelligence.

“When you ask someone from the vulnerability management team, ‘Hey, this server maybe is associated with the vulnerability, but is this server really exposed? And where is it in the network?’ He has no clue,” Noy said. “And the scanners don’t really even integrate with this network information, not to mention that even if they will pull it, because no pulling data is super easy with APIs.”

Noy said Astelia most frequently competes with Tenable, Rapid7, Qualys and CrowdStrike in the vulnerability management market. He’s confident in Astelia winning proof-of-value engagements since noise reduction is highly valued, network visibility creates differentiation and AI-driven exploitability analysis is novel.

“When you talk about infrastructure in large environments, you’re talking about firewalls and data centers,” Noy said. “Cloud doesn’t just set up networks and operate networks.”