Horizon3’s AI pentests spark security market shift

The gist
Horizon3’s NodeZero is rewriting the rules of cybersecurity by proving that fully autonomous AI pentesting is not just possible, but already dominating the market.
What to know
- NodeZero ran a production-safe lab test in late 2026, owning every domain and host in 14 minutes and 25 seconds, with zero humans in the loop.
- Horizon3 hit $100M ARR and raised a $250M Series E in 2026, riding a wave of demand for always-on, AI-driven security as attackers and defenders both turn to autonomous pentesting.
- The company is expanding fast, opening a Singapore ASEAN HQ and integrating NodeZero with Palo Alto’s Cortex XSOAR to make continuous AI red teaming a core part of everyday security ops.
AI Pentests Go Mainstream
NodeZero’s architecture and validation data proved autonomous pentesting had matured from lab demo to trusted operational tool, powered by billions of real-world attack and defense signals each day.
Autonomous pentesting mattered in late 2026 because it had already crossed from demo to dependable operating model. The AI Corner showed that NodeZero “was handed a range of IP addresses on a fully patched lab network with Microsoft Defender on every host,” then “owned every domain and every host 14 minutes and 25 seconds later, with zero humans in the loop,” while keeping production use credible through a design where real attacks are chained and fixes verified, with “every exploit” executed from validated code and models limited to planning and judgment. That architecture also fit the company’s framing of AI-native security as a system grounded in real offensive and defensive telemetry, including work associated with Berkay Celik and infrastructure “fed by an estimated 30 billion tokens of labeled attack-and-defense data a day.”
That technical maturity helps explain why late 2026 became an acceleration point rather than a speculative moment. NodeZero had already run “more than 350,000 production-safe tests across 7,000+ organizations,” and in the NSA Continuous Autonomous Penetration Testing program it logged 20,000 hours and found 50,000 vulnerabilities in its first year; against that backdrop, The VC Corner reported that “It passed $100 million in October 2026” and that “Eleven weeks before that milestone, the company closed a $250 million Series E,” with revenue still growing above 120% annually. Those milestones also aligned with strategic ASEAN expansion, reinforcing that the company’s late-2026 momentum was commercial and geographic as well as technical.
Threats And Funding Converge
Rising AI-powered attacks and a global investment boom are turning continuous AI security from a theoretical upgrade into a competitive necessity for both defenders and attackers.
The market is shifting because continuous AI-driven testing is no longer a conceptual upgrade but an economic and operational one: the September analysis on human-speed security argued that running “continuous AI red teams against your own estate” is cheaper than absorbing breaches “at machine speed with a human-speed budget,” while November reporting said security now “has to be always on.” That demand is landing amid a broader capital wave, with September research under the headline “Global VC Surge Fuels AI, Enterprise, Health, and Clean Tech Startups” citing Mistral AI’s “€3 billion (~$3.5 billion) Series D” and Cognition reportedly preparing to raise about $1 billion.
Adoption pressure is also coming from the threat side, where CrowdStrike documented an ARTEX campaign that “ran from late September to early October 2026 and ended with stolen data,” showing sustained operational use of AI pentest agents rather than one-off experimentation. CrowdStrike said the activity “let one operator target several financial institutions within days,” identified 38.244.50[.]120 as hosting “an ARTEX instance and an open directory” with a Claude Code prompt telling the model how to run tests, and found exposed directories with ARTEX configs, Claude session histories, and memory files in a two-server setup; the campaign “resulted in data theft,” including “customer names, phone numbers, annual income, loan limits, and borrowing-related data.”
Investors and Users Align
Horizon3’s rapid leap from $250M Series E to $100M ARR in weeks shows both capital and customers are betting big on autonomous cybersecurity as a market-ready solution.
The clearest proof that autonomous cybersecurity has reached meaningful commercial scale is that both capital markets and customers moved in quick succession. On 2026-08-03, Horizon3.Ai announced it raised $250 million in new funding to accelerate growth initiatives, a Series E that indicates investor confidence and provides a concrete signal of scaling momentum for its autonomous cybersecurity approach at a commercially meaningful level; by contrast, a September analysis about CrowdStrike is not usable here, because it contains no references to “Horizon3,” “Series E,” “August 2026,” “October 2026,” or “$100 million ARR.”
Customer validation arrived just weeks later in revenue form, showing this was not merely a financing story. Pulse 2.0 reported the headline, “Horizon3 Hits $100M ARR Driven by NodeZero AI‑Autonomous Continuous Pentesting, CEO Snehal Antani Says,” tying the milestone directly to the company’s autonomous product and indicating customers are adopting the model beyond pilots; taken together with the August raise, the October ARR crossing shows investors and buyers are already scaling the autonomous cybersecurity model at a size large enough to matter commercially.
Regional Strategy, Not Symbolism
Launching a Singapore HQ and hiring APJ leaders signals Horizon3’s intent to build deep, sustainable presence in Asia Pacific—not just chase headline deals.
The clearest sign that Horizon3 is moving beyond a single-market success is geographic institution-building: Yahoo Finance reported the headline, “Horizon3 Secures $250M Series E, Opens Singapore ASEAN HQ with EDBI and SG Growth Capital.” An ASEAN headquarters is not the language of opportunistic overseas sales; it signals a deliberate regional operating base, and the presence of EDBI and SG Growth Capital suggests the company is anchoring that base with local partners suited to sustained APJ and Southeast Asia go-to-market execution rather than treating Singapore as a symbolic outpost.
That interpretation is reinforced by how Horizon3 is staffing the region. IT Brief New Zealand’s headline, “Horizon3 Taps Chris Wood & Vivienne Horsfall in $250M Series E-Fueled APJ Expansion,” makes clear that Horizon3 is explicitly using its $250M Series E to expand its go-to-market in the APJ region, indicating growth is being operationalized beyond a single market; named leadership hires for APJ expansion show a company building repeatable regional coverage, management capacity, and field execution instead of relying on isolated customer wins to carry its international story.
Autonomous Pentests Meet Automation
By embedding NodeZero into Cortex XSOAR’s remediation workflows, Horizon3 is closing the loop between attack validation and automated response, making continuous Hack-Fix-Verify a reality.
NodeZero’s October 2026 launch into Palo Alto Networks’ Cortex XSOAR is significant less as a product add-on than as a workflow signal: autonomous pentesting is being wired into an incumbent security orchestration layer. As IT Brief New Zealand reported, the integration ties Horizon3 directly to Cortex XSOAR, and Horizon3’s own description makes the absorption explicit by saying XSOAR ingests NodeZero’s high and critical findings as incidents, then uses existing playbooks to enrich, route, track, and orchestrate remediation before sending teams back to retest.
That matters because Horizon3 explicitly frames the gap in current security automation as a lack of exploitability evidence and proof that fixes actually removed risk, then argues this integration closes that gap inside the established workflow. The company says teams often lack evidence showing what attackers can really exploit and whether remediation worked; by bringing pentest evidence into orchestrated remediation and pairing it with NodeZero’s 1-Click Verify, Cortex XSOAR becomes a continuous Hack, Fix, Verify, Repeat loop rather than a place where autonomous pentesting sits off to the side.










