AI Security

Adaptive Security made the Cyber 60 by deepfaking your CEO before a criminal does.

Built by the team behind Attentive, Adaptive maps what attackers can find about each employee, then runs personalized voice, text and video deepfake drills. Fortune reports $146 million. It claims 2,000+ customers. Now it wants the inbox too.

Nov 18, 2026 · 4 min read

The list. Adaptive Security made the 2026–2027 Fortune Cyber 60 in the early-growth group. Fortune’s profile lists New York City, CEO Brian Long, founded 2024, 290 employees. Fortune reports $146 million in funding, from Bain Capital Ventures, a16z, NVIDIA’s NVentures, the OpenAI Startup Fund and Capital One Ventures. That matches the company’s math: a $43 million Series A co-led by a16z and the OpenAI Startup Fund in April 2025, OpenAI’s first cybersecurity investment, later topped up to $55 million, then an $81 million Series B led by Bain Capital Ventures in December 2025. Total: $146.5 million. Valuation and revenue are Undisclosed. Lightspeed is not listed as an investor.

The team. Long and co-founder Andrew Jones previously built Attentive, the text-message marketing company, which Adaptive says grew to $500 million in annual revenue. That is a useful background here. They know exactly how to get a text read. Now they teach employees not to trust one.

The problem in plain words. Social engineering means tricking a person instead of hacking a machine. Adaptive cites figures that it is behind more than 95% of successful breaches and that deepfake incidents grew 17x from 2023 to 2024. The shift is in the channel. A few seconds of a CFO’s voice from an earnings call is enough to clone it, and the attack now arrives by phone, text or video call, places most security teams don’t watch. The old fix, a yearly training video plus a fake phishing email, was not built for that.

The one unique thing: the drill is built from your real exposure. Adaptive starts with open-source intelligence, or OSINT: what anyone can find online about your company and staff. Phone numbers. Headshots. Voice recordings. Exactly what an attacker needs. ISMG reported that Adaptive checks more than 1,000 OSINT data points per employee and plans to reach 10,000. Then it builds the simulation from that data. Per the same report, it can fake a live call from any colleague using about three seconds of audio, and make a video avatar from a couple of minutes of footage.

So the drill is not a generic “your package is delayed” email. It is your CEO’s voice, on your phone, asking for a wire. Training is then assigned based on how each person reacts. The scenario uses the same data a real attacker would, which means the test measures real risk, not quiz scores.

Why it is winning now. Speed. Adaptive says it went from public launch in January 2025 to more than 1,000 customers by March 2026 and more than 2,000 by September 2026, and that it hit that mark faster than Palo Alto Networks, Okta or CrowdStrike did. Company claims. Its Series B post names PayPal, Bose, the NHL, Figma and Perplexity among customers.

The expansion. In April 2026, Adaptive launched Agentic Email Security in beta. It connects to Microsoft 365 or Google Workspace through a read-only API, with no change to mail routing, and a team of AI agents checks every message. When an attack slips through, it is tied back to the employee it targeted, who gets assigned training. Adaptive says about half its customers use at least one email product, and that in one proof of concept it caught 80% more attacks than the customer’s existing tools. One test, company-reported. It also added AI governance, a browser extension that feeds shadow AI use into each person’s risk score.

The CyberMerge view. This is a land grab for “human risk.” KnowBe4 and Proofpoint own awareness training. Abnormal and the API email crowd own the inbox. Adaptive wants both by making detection and training one loop. Smart. Also risky. Email security is a knife fight, and Microsoft announced new defenses against synthetic media at Ignite this week.

Deepfake drills will not stay rare. Voice cloning keeps getting cheaper for every vendor, incumbents included. Adaptive’s edge has to be the data underneath, the exposure map and the per-person risk score, not the trick itself.

Bottom line. Adaptive is betting the human layer finally gets a real security product, not an annual compliance video. Watch how many customers add email, named enterprise wins outside tech, and whether a platform decides it would rather buy the human-risk score than build it.

Sources