AI Agent Control Is No Longer Optional: Why Onyx's $35M Reflects A Governance Crisis

There's a persistent myth in tech circles that AI agent security is a defensive feature—something bolted onto enterprise software after deployment, like a patch or plugin. The reality, revealed by Israeli startup Onyx Security's $35 million Series A in March 2026, is the opposite: autonomous AI agent governance has become operational infrastructure, demanded before agents go live.

When Onyx Security raised $35 million in its Series A led by Conviction, an AI-native venture capital firm founded by Sarah Guo, a former General Partner at Greylock, the capital flow signaled something deeper than another cybersecurity round. It validated that the fastest-growing problem in enterprise software is no longer who can access data—it's what autonomous agents do with it once they're inside.

Why The Misconception Exists

For years, cybersecurity teams sold control in binary terms: either you locked systems down, or you opened them. Traditional security meant permissions and audits. But when companies hand email access, file systems, or transactional databases to autonomous agents, neither approach works.

An agent doesn't request permission once—it executes thousands of operations inside workflows. It doesn't log into a system at a fixed time; it runs continuously. Traditional "user-based" controls assume human judgment at each decision point. Agents break that assumption.

Most enterprises initially treated this as a corner case: "We'll secure agents the same way we secure users." That framing is the myth. Onyx's system continuously detects AI assets, defines usage policies, and enforces them in real time, allowing AI deployment to scale safely—not because security teams wanted a new category, but because existing frameworks simply don't apply.

The Real Problem Onyx Solves

The misconception isn't that agent security matters. It's that companies believed they had time to figure it out gradually. They don't.

Onyx builds tools designed to detect, contain, and manage the unique failure modes that arise when autonomous AI agents are given access to corporate systems and data. The raise reflects investor concern about a fast-growing problem: as companies deploy agents to automate workflows—accessing email, file systems, scheduling, and transactional systems—those agents introduce new attack surfaces and reliability hazards.

This isn't theoretical. Onyx's technology secures more than 1.1 million agents for customers, and by July 2026, just four months after launch, the company raised an additional $113 million in a Series B, valuing it at $640 million. That speed reflects market validation: enterprises don't view agent governance as optional.

What Makes This Different From Earlier Cybersecurity Rounds

Israeli cybersecurity companies historically raised capital to defend against external threats—hackers, phishing, data exfiltration. Onyx raised capital to solve an internal problem: controlling the organization's own autonomous systems.

This is a category expansion, not a competitive iteration. At least 20 Israeli-founded cybersecurity startups are now competing to secure AI agents, as companies increasingly give autonomous software access to corporate applications, data and identities. But the funding disparity is telling: Onyx's $35 million Series A in March 2026 was followed by a $113 million Series B in July—a trajectory that reflects not competition, but market creation.

That market is real and quantifiable: the market for agentic AI security is projected to grow from $1.65 billion in 2026 to 13.52 billion by 2032. The growth rate (over 700% in six years) doesn't describe a feature category; it describes a new pillar of enterprise software infrastructure.

The Timing: Why March 2026 Mattered

Onyx didn't invent agent security in 2024 because agents were theoretical. The company launched when deployment crossed a threshold: enterprises stopped piloting agents and started embedding them into production workflows. At that inflection point, "we'll figure out governance later" became a business liability, not a technical delay.

Cybersecurity remained dominant but evolved in focus. Rather than traditional perimeter defense, many companies are targeting the risks created by autonomous systems and AI-driven workflows. Onyx Security raised $35 million to control AI agents within organizations, while Jazz secured $43 million to address data leaks in AI environments.

This marks a structural shift in Israeli cybersecurity fundraising. For two decades, Israeli cyber companies built detection and response tools. In 2026, they're building pre-deployment control layers—a preventative, not reactive, category.

Why Israeli Founders Led This Category

Israel's advantage in cybersecurity stems from military-grade operational thinking: systems that must work under adversarial conditions, with zero tolerance for failure. That mindset is directly applicable to autonomous systems, where a misconfigured agent can cascade into systemic risk.

Founded in 2024 by Maxim Bar Kogan and Gil Elbaz, the company launched publicly in March 2026 with $40M raised. Both founders brought expertise from systems that required absolute reliability—not post-hoc security, but built-in governance.

The Israeli startup ecosystem's concentration in AI-native infrastructure also mattered: the Calcalist 2026 list of the most promising startups shows a move toward "Steel-Tech" and autonomous systems, with agentic AI companies taking 24% of the list. That concentration allowed Israeli founders to see the agent security problem before it became mainstream.

How This Rebuts The Original Myth

What This Means For Israeli Tech In 2026

Israeli privately held technology companies secured $11 billion in funding during the first nine months of 2026, signaling a broad recovery. But the distribution wasn't uniform: large rounds in proven categories (cybersecurity, AI infrastructure) dominated capital.

Onyx's trajectory is the exception that proves the rule. It raised $35 million in March, $113 million in July—a 3.2x acceleration in four months—because it solved a problem that enterprises couldn't defer. Israeli venture capital is showing clear signs of recovery after two slower years, with Israeli funds raising nearly $900 million earmarked for new investments, with a further $600 million expected to be allocated in 2026.

As we covered in our analysis of Israeli High-Tech Sector Recovery as VC Investment Rebounds in 2026, capital is flowing back to Israel, but it's concentrated in companies solving infrastructure problems, not consumer-facing applications. Onyx exemplifies that pattern: it's not selling to end users; it's enabling enterprises to deploy AI safely. That's a multi-billion-dollar leverage point.

Key Takeaways for Decision-Makers

If you're an executive, technologist, or investor evaluating the Israeli startup ecosystem in 2026, Onyx Security illustrates three realities:

1. Categories Don't Grow At 700% For Compliance Reasons. The projected growth in agentic security reflects genuine operational necessity. Companies see agent deployment as competitive advantage, and they won't sacrifice speed to build custom security solutions.

2. Israeli Teams Understand Systems-Level Problems. Onyx's founders combined operator experience with technical depth—a profile common in Israeli tech but less common in Silicon Valley startups. That difference is becoming an advantage as AI systems require governance at the infrastructure level, not the application layer.

3. The Recovery Is Real, But Selective. Israel's tech recovery is real, but the money is becoming more selective, favoring companies with deep technology, strong infrastructure, and global scale. Onyx is a poster child: it raised $35 million to build infrastructure that enables a category (enterprise AI deployment) rather than serving one vertical. That selectivity will persist through 2026 and beyond.

FAQ: Six Questions About Onyx And Agent Security

Q: Did Onyx invent the problem, or did the problem already exist? The problem existed; Onyx saw it first. Bar Kogan and Elbaz said they teamed up to develop a platform to secure and control AI agents, as advanced AI systems are adopted faster than traditional enterprise security systems can keep up. Enterprise adoption outpaced existing governance frameworks. Onyx built the framework.

Q: Why did investors move so fast on the Series B? Four months between $35M Series A and $113M Series B is unusual. It signals confidence that the market will grow quickly, that Onyx can capture meaningful share, and that delays create competitive risk. When a VC firm commits $113 million in July, they're betting that by late 2026, agent security will be table stakes for enterprise AI deployment.

Q: Is this an Israeli phenomenon, or global? The market is global; Israeli leadership is factual. At least 20 Israeli-founded cybersecurity startups are now competing to secure AI agents. Palo Alto Networks CEO Nikesh Arora said recently that about 40 companies are currently being funded in the agentic security market, including a significant number of Israeli startups. Half the category leaders are Israeli or Israeli-founded.

Q: What does this mean for other Israeli cybersecurity companies? The market is expanding, not consolidating. Zenity raised $125 million in a Series C in August, bringing its total funding to approximately $185 million. Multiple Israeli companies are raising large rounds in agent security. That suggests the market is large enough for winners to coexist, not a zero-sum competition.

The Closing Insight

The original myth was that autonomous agents are a future problem requiring future security solutions. Onyx Security's $35 million Series A, followed by a $113 million Series B, proves the opposite: the future is now, the problem is present, and enterprises are funding solutions at venture-scale capital levels. For anyone tracking Israel's tech recovery, Israeli cybersecurity innovation, or the business case for AI governance, this is the most important startup narrative of 2026.

For more context on broader market trends, see our analysis of Iran War Escalation and Startup Resilience in Tech Fundraising, which examines how Israeli founders navigate geopolitical risk while scaling global companies.