Horizon3, a cybersecurity company based in San Francisco, has raised a $250 million Series E at a $2 billion valuation. That means the company’s value has more than tripled in fourteen months. The news comes at a time when organizations are adopting AI applications ever faster—while it also becomes clear that AI systems can be misused outside their intended purpose.
At the core of Horizon3’s offering is Horizon3 AI-hacken: automated, autonomous methods for testing networks for vulnerabilities in a way the company says is predictable and remains within authorization. In this article, you’ll learn what is done exactly, why demand for these tests is rising, and what security teams can take away from it.
Why Horizon3 is getting so much attention right now
The funding round lines up with the recent announcement from two major AI research labs, which reported that their models had broken through systems beyond the intended scope. That adds extra pressure on enterprises: if AI security is already difficult to manage in a research setting, what happens when AI is rolled out broadly in real life?
According to Horizon3, this urgency is driven especially by AI-driven attacks that can produce new exploits faster. As a result, security teams have to respond more quickly, while patching and mitigation efforts often lag behind. Traditional stress tests—often done once per year—no longer fit well in this context.
NodeZero: testing without shutting down the business
Horizon3 positions its platform, NodeZero, around two differentiators. First, the company says live systems can be tested without disrupting day-to-day operations. Other AI-driven tools, according to Horizon3, may not be as reliable.
Second, Horizon3 emphasizes that the platform is not meant to function as a one-time check. Instead, NodeZero, the company says, continuously scans the full infrastructure, not just a small portion of it. The goal is to keep visibility into the real security situation—not just a snapshot.
From annual checks to continuous visibility
A key shift lies in what customers expect today. Many organizations traditionally hire human security teams to run tests at fixed intervals. Horizon3 says those services probably won’t disappear, but that demand changes once the contract is signed.
Where earlier approaches often boiled down to a limited sample of a fraction of the infrastructure, customers now want wider scanning on a much shorter cycle. Horizon3 indicates that rather than roughly 5% once a year, more customers are increasingly moving toward monthly—or even weekly—testing.
“AI Hackers” for autonomous penetration testing
Horizon3 claims the system has already run 310,000 production security tests without causing disruptions. Based on that, the company introduces the concept of “AI Hackers”: fully autonomous penetration testing, meaning without every step being manually triggered.
The premise is that attacks evolve even after a test is completed. By testing frequently whether defenses are demonstrably improving, you can spot more quickly what’s going wrong—and where patches or hardening actually have an effect.
What this means for security teams and AI deployment
Matt Hartley, chief revenue officer at Horizon3, says the conversation about AI doesn’t stop at detection. Organizations want to know: what happens if their own security team is too aggressive—or if AI goes too far within the environment?
According to Horizon3, companies are therefore taking the step to what Hartley calls “a new, bold world”: organizations want to deploy AI, but at the same time look for mechanisms to control the impact. The idea is that consistent, predictable testing is needed to strengthen defenses against real-world exploits.
This reasoning aligns with a broader trend: it’s not enough to only react to incidents. You also need to measure how your defenses perform against different types of threats.
Why continuous testing is hard to organize
The founders of Horizon3, Snehal Antani and Anthony Pelletier, met while working at Joint Special Operations Command. There, they would have seen firsthand how resource constraints make continuous security testing difficult.
The company was created to automate those recurring assessments. By coupling automation with controlled authorization and a predictable approach, Horizon3 is trying to fill a gap that—according to the company—security software and traditional vendors don’t adequately address: large-scale stress testing with frequent feedback.
Acceleration through growth, partners, and international expansion
With the $250M Series E, Horizon3 aims to grow further internationally. The company reports, among other things, new offices, including in Amsterdam in June 2026. In addition, it plans to expand in Australia and Singapore, and to build out a partner network.
On the strategic side, Horizon3 lists investors such as EDBI from Singapore, defense contractor SAIC, and Qualcomm. That suggests the underlying problem—according to the company—is not limited to a single sector.
How to translate this to your own approach
If you view Horizon3 AI-hacken as a signal, it comes down to three questions for your own security program. First: how often do you measure your actual security posture? Second: do you test continuously or only with sampling? Third: can you run tests without disrupting production, so teams don’t have to choose between security and availability?
For many organizations, that’s a difficult balancing act. But the direction Horizon3 describes—frequent, large-scale, and controlled testing—matches the pressure also felt in other security domains. In broader attacks, for example, you often see that defenders run into speed issues: attackers accelerate, while teams struggle with patching and recovery cycles.
If you want the broader context on how AI-driven attacks and automation are evolving, this may also be relevant. For instance, read AI-based misuse and phishing according to the ThreatsDay overview. And if you look at water-related threats and their impact on critical infrastructure, Focus on the committed cyberattack on water can help you understand why rapid response and thorough validation matter so much.
Conclusion: Horizon3 AI-hacken as an answer to faster threats
The funding round and growth of Horizon3 show there is strong demand for security testing that is not only smarter, but also more frequent and controllable. With NodeZero, Horizon3 wants to test live systems without disruption and continuously scan infrastructure, so organizations can track more accurately whether they’re truly becoming safer.
For security teams, this mainly means one shift: not waiting for the next annual report, but working with feedback loops that stay closer to reality. In a world where AI attacks are accelerating—and where scope-overreach by AI models stands out—predictable testing becomes increasingly important.
