Mave, an Israeli startup building AI agents for security operations, announced $34 million in total funding on October 22, 2026. That includes a $28 million Series A led by Red Dot Capital Partners, with Lumir Growth Partners participating, and a $6 million seed round raised at founding. Dell Technologies Capital and Emerge Ventures led the seed and also joined the Series A.
Mave was founded in 2025 by CEO Raziel Oren and CTO Guy Eshel, both veterans of Israel’s Unit 8200. Its platform uses AI agents to investigate alerts across the security tools a customer already runs. The agents pull live context from that stack and combine fine-tuned small language models with frontier LLMs to reason through incidents in near real time.
Mave doesn’t promise fully autonomous response. According to SecurityWeek, its agents recommend actions, connect to tools with least-privilege access, add approval gates where needed and keep a full audit trail. “Autonomy without control is a risk,” Oren said. He described the goal as making AI “a responsible and accountable part of how the modern enterprise protects itself.”
The company has about 30 employees and plans to spend the money on AI research, its enterprise product and global expansion, including go-to-market hires.
Why it matters / what to do
Agentic SecOps is one of the busiest areas of security investment this fall, with a steady run of funding rounds for AI-agent security and autonomous security operations. For buyers, these products promise less alert fatigue and faster triage. The hard questions are about governance.
- Ask for proof of control: how agent permissions are scoped, which actions need human approval, and whether every agent decision can be exported for audit and compliance.
- Test on your own alerts: run a time-boxed pilot with your noisiest real alert sources, not vendor demo data, and measure investigation quality as well as speed.
- Plan the people side: decide in advance how analyst roles change if triage is automated, and keep humans in charge of containment decisions for now.
Source: SecurityWeek (Oct. 25, 2026)