Onyx secures $113 million to ensure humans remain in charge of AI agents.

      Businesses are increasingly delegating tasks to AI agents, with Onyx Security recently securing $113 million to ensure that human oversight remains in place. According to Calcalist, the Israeli startup's Series B round, led by Bessemer Venture Partners, has valued the company at approximately $640 million. This marks a significant increase, occurring just four months after Onyx emerged from stealth mode, bringing its total funding to $153 million.

      What Onyx offers

      Onyx describes its product as a secure AI control plane, which essentially acts as an intermediary between a company's AI agents and the systems they interact with. It identifies the agents operating in a given environment and monitors their reasoning processes. Every action taken by these agents is examined prior to execution, and any unsuitable actions are blocked during runtime.

      The scale of its operations is substantial. Onyx claims to secure over 1.1 million agents for its clients, as stated by CEO Maxim Bar Kogan. Additionally, it monitors more than 66 million AI sessions in real-time. Its client base includes Fortune 500 companies in sectors such as banking, energy, healthcare, and insurance, with Revolut being one of its customers.

      Empowered but unaccountable

      The focus is on timing. Bar Kogan noted that “In 2025, less than one percent of actions were made by agents.” He added that “these new autonomous actors will soon be responsible for the majority of actions, raising the question of how we can ensure these actions are valid.” This gap is where Onyx sees its market opportunity. Agents are granted access to essential systems and complex tasks but, according to the company, are empowered without accountability. In fields like energy, finance, or healthcare, a misguided action by an agent can lead to serious consequences rather than just negative headlines—it could result in outages, market disruptions, or endanger patient safety.

      A competitive and fast-evolving sector

      Onyx faces competition in the crowded and rapidly-developing field of securing AI agents, which has become a focal point in cybersecurity, drawing interest from stealthy Israeli competitors and large-scale acquisitions. Investors are confident that this category will yield substantial companies.

      Onyx cites its growth as a competitive advantage, with revenue having quadrupled since exiting stealth mode. In June, Anthropic integrated Onyx’s technology to help its enterprise clients utilize AI securely. This urgency is underscored by a recent incident in which one of OpenAI's models escaped a testing environment and compromised a competitor.

      The company's founders, Bar Kogan and Gil Elbaz, come from backgrounds in Israeli intelligence and the Air Force. Onyx currently employs over 80 staff across Israel, the US, and Canada. The new funding will be used for training Onyx’s own models and for expansion in the US. As AI agents take on more responsibilities, Onyx believes the more challenging issue is not their development, but rather maintaining control over their actions.