Israeli-founded cybersecurity company A Security emerged from stealth on Monday with $37 million in funding to expand its autonomous offensive security and remediation platform, the company announced.

The funding came from Lightspeed Venture Partners, Cyberstarts, Cerca Partners, and angel investors including Wiz CEO Assaf Rappaport and Cyera CEO Yotam Segev.

A Security said it would use the capital to scale its growth and expand a platform designed to discover and fix exploit paths before they can be used by malicious actors.

The company is entering the market as security teams face a new kind of threat from artificial intelligence-enabled attackers, who can search for weaknesses, chain them together, and automate parts of cyber operations far faster than traditional human-led attack teams.

"AI has changed the speed and scale of offensive cyber operations," said Yossi Torati, CEO and co-founder of A Security. "Attackers are using agentic capabilities to find, chain and exploit weaknesses across environments, at a pace that human teams relying on manual processes cannot match."

Torati said the company was built to discover "real-world exploit paths" that AI-enabled attackers would use to cause breaches, downtime, and business disruption, and help organizations remove them before they are exploited.

Before founding A Security, Torati spent nearly two decades working with global enterprises on cyberattacks and resilience, most recently as director of enterprise security at incident response firm Sygnia.

The company’s other co-founders, Omer Gull and Yuval Itzchakov, bring experience from Check Point Software, Hunters, and the IDF’s Unit 8200, according to the company.

A Security said its platform differs from periodic penetration tests and manual risk assessments by continuously identifying cross-domain attack paths and validating whether they can actually be exploited.

The system uses offensive and defensive AI agents to test environments under scoped execution and full audit trails, then helps drive remediation at the source and through other security controls, the company said.

"Today’s cyber defenders need a weapon that can beat weaponized AI to the punch," said Guru Chahal, partner at Lightspeed. "We believe A Security is that platform."

Hila Zigman, general partner at Cyberstarts, said the security industry was underestimating how dramatically AI was changing offensive cyber operations.

"Attackers are already operating with autonomous capabilities, at a speed and scale defenders were never designed to handle," Zigman said.

The company said it is already working with large organizations in finance, healthcare, critical infrastructure, and technology, where security teams are under pressure to validate and close attack paths faster than traditional testing cycles allow.

The launch comes amid growing concern in the cyber industry over AI systems that can assist in vulnerability discovery and exploitation. Reuters reported last week that Anthropic was expanding access to Claude Mythos Preview, an AI system designed to help identify software vulnerabilities, to about 200 organizations through its Project Glasswing initiative.

A Security said its platform is aimed at organizations "where failure is not an option," including sectors that cannot afford operational disruption from cyberattacks.