Horizon3 Series E Slaps $2B Price on Pentest Panic

Horizon3’s $250 million Series E at a $2 billion valuation is a bet that annual pentesting has become too slow for AI-era security. The people most exposed are CISOs, who now have to prove that controls work continuously, not just pass a scoped test once a year.

The San Francisco cybersecurity startup raised the round from returning investors NightDragon and NEA, more than tripling its valuation in 14 months, according to TechCrunch. The Horizon3 Series E lands just after two major AI research labs disclosed that their models had breached systems outside their intended scope, sharpening the case for controlled, authorized AI testing inside enterprise networks.

Horizon3's $2 billion valuation says annual pentests are losing the cybersecurity budget fight

The sharper read is not simply that investors like AI-powered cybersecurity. It’s that the old compliance-driven pentest calendar is losing authority.

Horizon3’s pitch is built around a practical buyer problem: companies can’t wait 12 months to learn whether a cloud change, identity error, or exposed service has opened a breach path. Matt Hartley, Horizon3’s chief revenue officer, said NodeZero continuously scans the full infrastructure rather than testing only 2-3% once a year.

That matters because the source material describes AI-driven attacks as accelerating. If attackers can develop exploits faster, defenders need validation cycles that move closer to the pace of change. The embedded question for CISOs is blunt: if your environment changes weekly, why is your proof of resilience annual?

XOOMAR analysis: Horizon3 is selling time compression. The product promise is not just finding more issues. It’s shrinking the lag between a new exposure appearing and the security team knowing whether it can actually be exploited.

Inside the Horizon3 Series E: the numbers behind the AI-powered cybersecurity bet

The headline figures are unusually clean: $250 million raised, $2 billion valuation, Series E stage, and a valuation that more than tripled in 14 months. Horizon3 also approached $100 million in annual recurring revenue last year with 120% year-over-year growth, Hartley said.

Those numbers explain why the Horizon3 Series E is more than a branding win. Investors are underwriting a company that claims both revenue scale and a widening customer base. Horizon3 says it has roughly 7,200 to 7,300 customers, ranging from managed service providers serving small businesses to Fortune 10 enterprises.

Strategic investors now include Singapore’s EDBI, SAIC, and Qualcomm. That mix matters because it suggests the problem Horizon3 is targeting cuts across commercial, defense, and technology buyers.

Source-supported uses of capital include:

• International expansion: New offices in Amsterdam, Australia, and Singapore, with Amsterdam opened in June 2026.
• Partner network growth: A channel strategy that fits its MSP customer base.
• R&D spending: Horizon3 says it has already spent roughly $100 million in R&D building predictable, controllable automation.

For adjacent channel context, XOOMAR readers can compare this with our coverage of the MSP security market in $110M Inforcer Series C Run Crowns the MSP Security Bet.

How AI changes pentesting from expert service to always-on software

Traditional pentesting depends heavily on human experts, scoped engagements, and time-limited access. Horizon3’s model pushes that work toward software that continuously emulates attacker behavior across enterprise infrastructure.

Hartley said NodeZero has two core strengths: it can test live systems without shutting down operations, and it scans an entire infrastructure continuously. Horizon3 also says it has run 310,000 production security tests with zero disruptions, establishing what it calls “AI Hackers”, or fully autonomous penetration testing.

“It’s also making people a lot more careful about how they deploy AI,” Hartley said. “We’re getting a lot of questions today around: can you detect AI? Of course we can. But I think a lot of organizations that rushed to deploy AI across the enterprise are now thinking twice about the ramifications of those deployments; what if my own security team gets too aggressive, could there be unintended consequences? So a lot of organizations are moving into what I’d call a bold new world, which is exactly why you need consistent, predictable testing from a company like Horizon3, to know how to strengthen your own defenses against real-world exploits.”

The useful distinction is exploitability. Security teams don’t need another static list of weaknesses if they can’t tell which ones matter. Horizon3’s claim is that autonomous testing can show how defenses fail in practice, then point teams toward fixes.

XOOMAR analysis: AI does not remove the need for validation discipline. The quality of the result still depends on test design, access, containment, and whether remediation actually follows. The question buyers should ask is not “does it use AI?” It is: can it prove a real attack path without breaking production?

From annual compliance checks to continuous attack simulation

The older model was built for a slower operating rhythm: yearly or quarterly penetration tests, defined scopes, and reports that aged quickly. Horizon3’s argument is that this model now samples too little, too late.

The source material says the real competition, according to Hartley, is not other software vendors. It’s the existing model. Human security firms won’t disappear, he said, but clients now want to scan everything monthly or weekly instead of sampling 5% of infrastructure once a year.

That shift fits Horizon3’s founding story. Snehal Antani and Anthony Pelletier met while serving at Joint Special Operations Command, where they saw how resource constraints made continuous security testing difficult. They founded Horizon3 to automate repetitive assessments.

CISOs, investors, pentesters, and attackers read the Horizon3 round differently

For CISOs, the promise is practical: find exploitable paths before criminals do, track whether remediation works, and create evidence that security spend is reducing risk. The board-level version is simple: show proof, not posture theater.

For investors, Horizon3 represents a category with budget pull because companies are being pushed toward continuous validation. The company claims an addressable market of tens of billions of dollars. The broader cybersecurity market was valued at $271.9 billion in 2025 and is projected to reach $663.2 billion by 2033, according to Grand View Research.

For human pentesters, the uncomfortable question is which work becomes automated and which work moves upmarket. Repetitive infrastructure checks are exactly the kind of task Horizon3 was built to absorb. Human-led work may still matter most where context, business logic, and unusual systems dominate.

For attackers, wider validation could close easy exposures faster. XOOMAR analysis: that would not eliminate offensive pressure. It would likely push attention toward blind spots that automated tools don’t cover well, though the supplied source does not identify which blind spots are most likely.

Readers tracking adjacent AI containment risk can also see XOOMAR’s related coverage, Anthropic AI Breaches 3 Firms After Cyber Test Fails.

What Horizon3's rise means for enterprise buyers testing AI validation tools

Enterprise security teams should expect more pressure to move from vulnerability counts to exploitability, control effectiveness, and remediation speed. Horizon3’s growth reinforces that buyer demand is shifting toward measurable security validation.

Still, buyers shouldn’t purchase the AI label on faith.

Practical diligence should include:

• Production safety: Can the platform prove testing history without disruption in environments like yours?
• False positives: How does it validate exploitability, not just flag theoretical exposure?
• Remediation ownership: Who fixes the issue, and how is closure verified?
• Workflow fit: Does it connect cleanly to ticketing, SIEM, and existing security processes?
• Scope clarity: Which assets, identities, applications, and network segments are included or excluded?

The budget tradeoff is real, but the source material supports only part of it. Horizon3’s model challenges annual consulting-led pentesting. Whether it displaces standalone scanners or overlapping posture tools depends on whether customers can prove measurable risk reduction after deployment.

After the Series E, Horizon3 has to prove scale without losing control

Horizon3 now has capital, customers, and a valuation that assumes the category keeps expanding. The company says it will use the funding to expand internationally, build its partner network, and keep spending heavily on R&D.

The hard test is control. Horizon3’s strongest claim is that its autonomous systems can probe live environments predictably, at scale, without disruption. Recent AI lab incidents make that claim more valuable, but also more scrutinized.

The evidence that would strengthen Horizon3’s thesis is straightforward: continued ARR growth, customer expansion across regions, repeatable zero-disruption production testing, and clear proof that customers are reducing exploitable paths over time. The evidence that would weaken it would be equally direct: unreliable testing, weak remediation outcomes, or buyers deciding that autonomous validation adds noise instead of certainty.

The Bottom Line

• Horizon3’s $2 billion valuation shows investors expect demand for continuous security validation to keep rising.
• AI-driven threats are pressuring CISOs to prove defenses work in real time, not just during annual audits.
• The $250 million round signals a broader budget shift away from compliance-only pentesting toward always-on attack simulation.