Via themedialine.org
The funding round reflects a broader wave of capital flowing into companies building defenses for autonomous AI systems
An Israeli cybersecurity startup has secured $140 million in fresh funding aimed at protecting AI models from misuse, arriving at a moment when high-profile AI hacks have made the vulnerability of autonomous systems impossible to ignore.
A surge of capital into AI defense #
Torq, one of the more prominent players in this space, raised approximately $140 million in a Series D round in January 2026, part of a broader $332 million funding haul that pushed its valuation to $1.2 billion.
Torq’s focus is on Security Operations Centers, or SOCs, that use autonomous agents. Think of it as building AI to watch AI, automating the kind of threat detection and response work that human security analysts currently handle, but at machine speed and scale.
Torq isn’t alone. Oligo, another Israeli startup specializing in runtime security for AI agents and cloud-native applications, reached $140 million in total funding after closing a $60 million round in August 2026. Onyx Security pulled in $113 million in July 2026, bringing its total to roughly $153 million, for work on what it calls an “AI control plane” designed to manage and govern autonomous agents. And Irregular, which collaborates with OpenAI and Anthropic on red teaming strategies, raised around $80 million in 2025.
Red teaming, for those unfamiliar, is essentially the practice of hiring people (or building systems) to deliberately attack your own AI to find weaknesses before bad actors do.
The competitive landscape and what to watch #
What’s notable about the cluster of companies attracting this capital is the diversity of approaches. Torq is building automated SOC infrastructure. Oligo is focused on runtime security, essentially monitoring AI agents while they’re actually running rather than just testing them before deployment. Onyx Security is tackling the governance layer, trying to create centralized control systems for managing fleets of autonomous agents. Irregular is working upstream, partnering with the labs building frontier models to probe for vulnerabilities before they reach production.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our