Israeli company Mave announced on October 22, 2026 that it was emerging from stealth with total funding of 34 million dollars for an AI agent platform connecting an organization's existing security tools. That total includes a new 28 million dollar Series A led by Red Dot Capital Partners with participation from Lumir Growth Partners, alongside an earlier 6 million dollar seed round led by Dell Technologies Capital and Emerge Ventures. The seed investors also joined the new round.
Its stated aim is to help security teams investigate alerts, hunt for suspicious activity and execute approved responses without first moving all organizational data into a new centralized database. The company says its platform is commercially available. The funding will support AI development, the enterprise product and international expansion. The announcement therefore combines financing with a product already offered to customers, rather than only a plan for future development.
Connecting evidence where it already lives
A security investigation may require information from an identity platform, endpoint protection and a cloud environment. One system might show an unusual login, another activity on a computer, and a third a permissions change. This example illustrates the work Mave aims to bring together: gathering context from different sources, connecting events and explaining their relationship. It is an illustrative workflow, not an account of a particular customer incident.
According to Mave, its agents query connected tools in their native query languages and adapt investigations as evidence emerges. An alert, a team member's question, a new attack campaign or a detection gap can provide the starting point. Data is retrieved for the task from its existing location, rather than requiring the organization to copy all telemetry into another repository first. Direct system access still requires suitable integrations and permissions.
That distinction matters to buyers. Removing a full data migration requirement does not mean operating without access to information. The platform must read the necessary data and stay within boundaries set by the customer. The scope of accessible information, its permitted uses and each connection's permissions belong in the deployment decision. They cannot be inferred simply from a promise to connect existing security tools.
More than a chat window over a security product
Mave describes a workflow spanning threat intelligence, threat hunting, investigation, detection engineering and response. A new attack campaign, for example, could prompt an assessment of its relevance to the organization's technologies and identities, a search for related activity and an investigation of findings. Where justified, the system can help create a detection or a response permitted by policy. These are company descriptions of capabilities, not a promise that every task will produce the same outcome.
Mave's distributed announcement says the product combines purpose-built workflows, adapted small language models and advanced large language models. The division is intended to support both repetitive work at high volume and complex reasoning. The announcement does not provide a complete model list, a cost per investigation or a comparative benchmark allowing buyers to calculate an advantage over human processes or competing products. The funding itself therefore provides no basis for claiming technological superiority.
Who authorizes actions on real systems?
When an AI agent moves from recommending to acting, the question is not only whether it identified a threat correctly, but what it may do next. Mave says the security team decides which actions can run automatically and which require human approval. It describes recording queries, evidence, decisions and actions so staff can inspect how a conclusion was reached and why a response was proposed.
The company says the platform has already helped during live incidents by isolating systems, disabling compromised identities and blocking malicious activity. The announcement lacks detailed incident accounts explaining its behavior in each case. These capabilities require a distinction between an action the agent proposes and one it is authorized to execute. Permission to isolate a computer or disable an account can also affect legitimate work, making approval policy central to the product.
Commercial figures and the Israeli connection
Mave was founded in 2025 by CEO Raziel Oren and CTO Guy Eshel, who met during service in Unit 8200. According to the announcement, Oren previously worked on go-to-market activities at Wing Security, while Eshel was Alterya's first employee and later its vice president of R&D before its acquisition by Chainalysis. The funding release says Mave employs more than 30 people and plans further hiring in AI and commercial roles.
On its website, Mave says it operates in more than 200 security operations centers, or SOCs. That is a company-reported count of centers, not necessarily the number of direct customers or separate organizations. It describes working with managed security service providers and large enterprises. Usage figures and success accounts come from the company; they indicate reported commercial activity but are not independent measures of accuracy, coverage or reliability.
What the funding makes possible to assess next
For an enterprise customer, the practical question is whether integrations provide enough context to reach a useful decision. Alongside investigation time, evaluations should examine evidence quality, incorrect conclusions, human interventions and the ability to stop or correct an action. These are assessment criteria for this type of product, not results published by Mave. Pricing, integration maintenance and staff training will also influence total deployment cost.
The announcement contains neither a price list nor a complete list of supported systems. Prospective customers must establish which connections support their environment, who maintains them and the permissions required for each action. Commercial availability does not guarantee automatic compatibility with every existing infrastructure.
The funding gives Mave resources to expand its product and operations, and the announcement sets out a clear approach: work across existing systems while leaving action policies with the security team. The commercial and technical test will be preserving that approach as systems, users and investigations increase. For Israel's AI industry, this is new funding for a product that extends beyond generating answers into organizational actions with tangible consequences.