The list. Prophet Security made the 2026–2027 Fortune Cyber 60 in the early-stage group. Fortune’s profile puts it in Palo Alto with CEO Kamal Shah, founded in 2023, 90 employees. Fortune reports $41 million in funding, from Bain Capital Ventures, Accel, Citi Ventures and Amex Ventures. That lines up with what the company has announced: an $11 million seed led by Bain Capital Ventures in April 2024, and a $30 million Series A led by Accel, announced July 29, 2025. Amex Ventures and Citi Ventures added strategic money in February 2026, amount not given. Valuation and revenue are Undisclosed. Lightspeed, which helps run the Cyber 60, is not listed as an investor. Prophet says this is its second year in a row on the list.
The problem in plain words. A security operations center, or SOC, is the team that watches alerts from every security tool and decides which ones are real attacks. There are too many alerts and not enough people. Prophet’s own survey of 250 IT and security pros, run by ViB, found teams leave an average of 28% of alerts uninvestigated. And 40% said they had disabled, or thought about disabling, detection rules because they couldn’t keep up with the alerts those rules produced. Read that again. Teams are switching off their alarms because the alarms are too loud. Company-sponsored survey, so treat it as directional.
What Prophet does. Its AI SOC Analyst picks up each alert, makes an investigation plan, pulls evidence from the tools a customer already runs (EDR, identity, cloud, email, SIEM), and writes up a verdict in plain English with a timeline. A human can see every question it asked and why it decided what it did. Accel’s partner called “showing their work” the differentiator when it led the Series A. Prophet also sells an AI Threat Hunter and a 24/7 human backup service called Watchtower.
The scale claims are company figures. Prophet’s Q3 2026 threat report says its AI analyst asked 4.7 million questions across a full quarter of alerts, a median of 35 per investigation, and reached a median verdict in just over five minutes. 93% of finished investigations were benign and 7% malicious. That last split is the whole SOC problem in one line: most of the work is proving that nothing happened.
The one unique thing: coverage graded by real investigations. In July 2026 Prophet shipped AI Detection Engineer. Most “detection coverage” dashboards count which rules are switched on and map them to MITRE ATT&CK, the standard catalog of attacker techniques. Prophet builds its ATT&CK map from completed investigations and threat hunts instead. So it can show which techniques your rules actually catch, which rules have gone quiet, and where the blind spots are. Then it drafts new detections in your existing stack, tunes noisy ones, and backtests every change against your own history before anyone approves it. Every change is version controlled and can be rolled back. Customers choose how much it can do alone.
Why that matters: it closes a loop most SOCs never close. Investigations normally end in a ticket. Here they become training data for the rules that create the next alerts. Fewer junk alerts means fewer reasons to switch rules off. Macnica, which now resells Prophet in Japan, describes the same loop: detection findings sharpen investigations, and investigation results sharpen detection.
Why it is winning now. The money is moving to AI-run security operations. This morning, Quorum Cyber closed its purchase of Ontinue, an AI-powered managed security operations provider, to build what it calls autonomous cyber defense. Terms Undisclosed. Prophet sells software to in-house teams that might otherwise renew an MDR contract (outsourced 24/7 detection and response). It has written openly about that renewal question. Partnerships widen the funnel: Google Security Operations, ExtraHop’s network data at no extra charge, and Macnica in Japan.
The CyberMerge view. Triage speed is becoming table stakes. Dropzone AI, Torq, Simbian and the big platforms all sell AI alert triage now. A five-minute verdict is a good demo, not a moat.
The moat is the loop. A vendor that sees every investigation outcome knows which detections are dead weight. That data is hard to copy and it compounds. It also moves Prophet up the org chart, from “help the tier 1 analyst” to “tell the detection engineering lead what to fix.” That is a budget line with a director’s name on it.
The risks: Microsoft, Google and CrowdStrike can bundle a good-enough AI analyst into the SIEM or EDR a customer already pays for. And every efficiency number here is self-reported. Prophet needs named reference customers on detection outcomes, not just speed.
Bottom line. Prophet is betting that the AI SOC’s real job is not closing alerts but fixing the rules that create them. Watch for published before-and-after alert volumes, more named customers like ETS, and whether MDR providers start buying this kind of engine rather than building it.
