Horizon3.ai

A little over a year ago, I took a close look at Horizon3.ai before deciding to join. I was convinced by what I saw then. One year later, even more so. In just over a year, we’ve tripled our valuation to more than $2B. We have a fantastic product, strong market momentum, an incredible team, and we’re growing fast. And what a year it has been in EMEA. Our team has grown almost 180%, we’ve welcomed many new partners and customers, and in Amsterdam we managed to outgrow two offices before finally finding our new home. A big thank you to everyone who made this first year such a good one. My team, my leadership team, everyone across Horizon3.ai, our partners and our customers. And to those who trusted me with the opportunity in the first place, thank you. I’m very glad I said yes! What a first year. Excited for what’s next!

Great to see Ayscom and Horizon3.ai bringing this conversation to the market. After a very positive meeting with the Ayscom team today, I’m excited about the plans we are building together and the momentum ahead. This webinar is a great opportunity to dive into AI, Mythos and how security teams can better prioritize exploitable risk. Muchos gracias Andres Sevillano Castaño and Francisco Javier Martín Ramos!

Last Wednesday, we officially opened our new EMEA HQ in Amsterdam. A proud moment for Horizon3.ai, and a very special evening with our founders and their families, our executive leadership team, colleagues, and the local team in the room. There was one thought from the evening that really stayed with me: when you start a company with an idea, a vision, and a lot of ambition, seeing your company name on the wall in another country, across the ocean from where it all began, must be incredibly special. And it was! A big thank you to everyone who invested in and helped create our new home, and to everyone who joined us to celebrate this next chapter for us in EMEA. A proud milestone for Horizon3.ai!

Seven years ago, Horizon3.ai saw where cybersecurity was heading and built NodeZero for the era we’re entering now. As threats like Mythos accelerate real-world attacks with AI, proof-based security validation is becoming essential to help organizations stay ahead of what attackers can actually exploit. #Mythos #ProjectGlasswing #AutonomousPentesting #OffensiveSecurity #NodeZero

Per Verizon DBIR, 85% of CISA KEV’s are still unpatched 30 days after KEV designation. A customer who is in beta with one of the #projectglasswing vendors found 1m+ vulnerabilities with 400k ‘exploitable’. There are 50 companies with early access to Project Glasswing. Open AI’s #daybreak launched yesterday with eight primary cyber vendors and is designed to scale through #trustedaccessforcyber (TAC). Board is demanding a plan NOW and a solution in place by September. Three important recommendations that Horizon3.ai can support with: 1. Enhance responsiveness to CISA KEV disclosures by reducing reaction and remediation timelines from days to hours. 2. Strengthen vulnerability management programs by prioritising remediation of exploitable vulnerabilities known to be used by threat actors aligned to precise business impact 3. Improve security controls to reduce attacker dwell time while simultaneously advancing automated remediation capabilities, including #precisionpatching , to accelerate containment and recovery. #mythos #cyberresilience #gohackyourself

The future (current) cyber fight isn’t about Humans vs. AI. It’s AI vs. AI, with humans deciding who wins. - Attackers are scaling faster than ever - Mythos, GPT 5.5, name the next publicly available frontier model! - Autonomous exploitation, AI-assisted recon, and the coming #Vulnpocalypse are going to stress every security team on the planet. - The answer won’t just be more hardware, more dashboards, or more alerts - you need the best AI on your side, enabling your people to ruthlessly prioritize winning even among a Tsunami of vulnerabilties. SOF Truth #1 still applies: “Humans are more important than hardware.” The organizations that win will combine great people with AI-enabled defense, rapid decision making, and teams built on trust, adaptability, and execution. Come grab a drink and build trust with the builders, defenders, and innovators who are deploying the future of cyber warfare today at Horizon3.ai with NodeZero. Happy Hour details here: https://luma.com/n92i01v4 SOFWEEK BOOTH: 2421 #CyberSecurity #AI #Vulnpocalypse #ExposureManagement #CTEM #SecurityOperations #SOC #ThreatExposure #Leadership #SOFTruths #CyberDefense #SecurityEngineering #SOFWEEK

Now that May has started, it feels like a good moment to look back on a very busy April for Horizon3.ai in Amsterdam: • We officially established our Dutch entity • Amsterdam became our EMEA HQ • We moved into our new office on the Herengracht, arguably the most beautiful canal in Amsterdam 😄 • We hosted bootcamps for 90 technicians from our partner community • We welcomed 12 new hires during our 3rd EMEA bootcamp, another great example of the continued growth across EMEA • We brought cross functional teams together for planning sessions • And we closed our first fiscal quarter with Amsterdam already operating as a true regional hub What I like most is that the office is already doing what we hoped it would do: bringing people together. Customers, partners, new hires, colleagues and teams from across the business. A big thank you to everyone who helped make this happen. April was a strong start. Excited for what’s next! #CyberSecurity #Amsterdam #EMEA #Growth #CompanyCulture

Hackers using LLM-based AI Agents will NOT be a threat to enterprise networks. It’s a bold opinion, but hear me out … LLM’s bias towards actions they’ve “seen” before. We know that Offensive AI agents are powered by LLM’s that are typically built on training data from in HackTheBox, TryHackMe, VulnHub, and other public CTFs That means agents will have a bias for pursuing actions and attack paths they’ve seen before So while these agents demo well, their training data is actually a huge weakness similar to model poisoning, where these agents will “chase” attack paths if you bait them the right way And remember, the goal is to generate enough signal to trigger defenses to contain and eradicate Tl;dr: Enterprises that deploy the right type of deception can completely undercut offensive AI agents built on LLM’s #ai #infosec #cybersecurity #infosec Horizon3.ai

Pleased to see this momentum reflected in our FY2026 results at Horizon3.ai! With 102% ARR growth year over year, 125% net dollar retention, and a growing global footprint across more than 5,200 organizations, the demand for AI-driven offensive security and real-world validation continues to accelerate. #offensivesecurity #security #cyber #ai #offense