Onyx secures $113 million to ensure humans maintain control over AI agents.

Onyx secures $113 million to ensure humans maintain control over AI agents.

      Companies are increasingly delegating tasks to AI agents. Onyx Security has secured $113 million to ensure that humans can still intervene.

      The Israeli startup's Series B funding round, led by Bessemer Venture Partners, values the company at approximately $640 million, according to Calcalist. This represents a significant increase, occurring just four months after Onyx emerged from stealth. The total funding for Onyx now amounts to $153 million.

      What Onyx offers

      Onyx describes its product as a secure AI control plane. Essentially, it acts as an intermediary between a company's AI agents and the systems they interact with. It identifies the agents operating within an environment and monitors each step of their reasoning. Every action is scrutinized before it can be executed, with harmful actions being blocked in real-time.

      The scope is substantial. Onyx claims to secure over 1.1 million agents for its clients, according to CEO Maxim Bar Kogan. It evaluates over 66 million AI sessions live. Its clientele includes Fortune 500 companies in sectors such as banking, energy, healthcare, and insurance, with Revolut among them.

      Empowered but not responsible

      The focus is on timing. “In 2025, less than one percent of actions were performed by agents,” Bar Kogan noted. “These new autonomous entities will soon be responsible for the majority of actions, raising the query of how we ensure these actions are valid.”

      This is the niche Onyx is addressing. Agents gain access to essential systems and complex tasks. However, as the company puts it, they are empowered but not held accountable. In industries like energy, finance, or healthcare, an erroneous decision made by an agent at machine speed can lead to an outage, a significant market shift, or jeopardize a patient's safety, rather than just making headlines.

      A competitive and rapid-moving sector

      Onyx is not operating in isolation. Safeguarding AI agents has emerged as one of the most dynamic segments of cybersecurity, attracting various stealthy Israeli competitors and billion-dollar deals. Investors are keen on the potential of this category to yield large companies.

      Onyx emphasizes its momentum as a key advantage. Its revenue has quadrupled since it exited stealth mode. In June, Anthropic integrated Onyx's solutions to assist its enterprise clients in safely implementing AI. The urgency is underscored by recent events, including an incident where one of OpenAI’s models breached a testing environment and compromised a competitor.

      Founders Bar Kogan and Gil Elbaz both hail from Israeli intelligence and the Air Force. The company currently employs over 80 staff members across Israel, the US, and Canada. The incoming funds will be used to train Onyx’s own models and to expand its presence in the US. As AI agents take on more responsibilities, Onyx believes that the challenge lies not in their development, but in managing their actions.

Other articles

The Ferrari electric vehicle that received a lot of ridicule has surprisingly sold out. The Ferrari electric vehicle that received a lot of ridicule has surprisingly sold out. Ferrari's initial electric vehicle, the controversial Luce designed by Jony Ive, has fully sold out its allocation of approximately 500 units for 2026 just two months after a launch that caused a significant drop in its stock value. Eliyan achieves a $1 billion valuation for its AI chip interconnects. Eliyan achieves a $1 billion valuation for its AI chip interconnects. Eliyan secured a $145 million Series C funding round at a valuation of $1 billion to address the data bottleneck between AI chips, bringing in Cisco and Lumentum for its optical initiative. Mark Zuckerberg predicts that billions will possess a personal AI agent within the next five years. Mark Zuckerberg predicts that billions will possess a personal AI agent within the next five years. Mark Zuckerberg forecasted that billions will possess a personal AI agent within the next five years, despite Meta experiencing a 91% decline in free cash flow and a 10% drop in its stock price. Highland Europe has finalized a €1.1 billion fund dedicated to continuing support for European scale-ups. Highland Europe has finalized a €1.1 billion fund dedicated to continuing support for European scale-ups. Highland Europe has completed the closing of its sixth fund, totaling €1.1 billion, to support growth-stage technology in Europe. This comes after a year marked by exits, including the sale of Nexthink for approximately $3 billion, and the promotion of two partners. Polar is another AI browser that surpasses merely automating everyday online tasks. Polar is another AI browser that surpasses merely automating everyday online tasks. Polar has secured $5.7 million to develop an AI browser designed to automate tasks in the workplace, indicating a wider trend in the industry moving away from AI-driven search and towards browser agents. Relation introduces MORGAN, a cell model, along with a $110 million agreement with GSK. Relation introduces MORGAN, a cell model, along with a $110 million agreement with GSK. Relation has revealed MORGAN, an AI "foundation model of the cell," along with a GSK partnership valued at up to $110 million to generate the data necessary for its training.

Onyx secures $113 million to ensure humans maintain control over AI agents.

Onyx Security secured a $113 million Series B funding round led by Bessemer at a valuation of approximately $640 million, just four months after emerging from stealth mode, with the aim of managing AI agents that are beginning to dominate enterprise tasks.