The Tel Aviv and New York startup builds artificial attack evidence to test security systems without executing real attacks.
Brava Security has emerged from stealth with $15 million in seed funding led by Pitango and F2 Venture Capital.
The November 4 announcement also disclosed participation from individual investors associated with Anthropic, Microsoft and Check Point. Brava is developing a platform that evaluates a security organization's ability to detect attacks before those attacks happen.
Simulating the Traces Attacks Would Leave
Many enterprises invest heavily in security information and event management systems, but an existing detection rule is not proof that an attack will trigger an alert.
Logs can be missing, event fields can change and telemetry pipelines can fail. Brava's approach uses AI agents to model attacks and synthesize the evidence those attacks would generate in a customer's specific environment, without executing the attacks themselves.
The system checks whether existing detection tools would recognize the resulting signals, identifies gaps and helps rewrite failing detection rules. Brava also offers an investigation agent that queries available security data and links its findings to underlying events.
According to the company, it processes approximately 2.4 trillion events per day across customer environments and replay-tests more than 1,100 attack techniques. Those numbers are company-reported operational metrics, not independent validation of detection accuracy.
Founders From Microsoft and Israeli Security Research
Brava was founded by CEO Idan Basre and CTO Itay Gabbay in 2024.
Basre previously worked on cybersecurity research in Israel's Unit 8200 and product management for Microsoft Defender. Gabbay worked on AI systems in the Israeli military, including joint research associated with DARPA.
The startup says it already works with Fortune 100 companies in retail, energy and hospitality, as well as other large organizations. It integrates with tools including Splunk, Google SecOps and CrowdStrike.
The Accountability Layer for Autonomous Security
As enterprises entrust more security operations to AI agents, missing evidence becomes a greater risk: a model cannot investigate a threat it never observes.
Brava aims to sell continuous verification that detection systems actually work. The seed funding will support research, product development and expansion across North America.
The company has not disclosed revenue or customer contract values. Its commercial challenge is showing that simulated attack evidence predicts real-world detection performance sufficiently well for security teams to rely on it.
Sources: Brava launch announcement · F2 Venture Capital · CTech