NATO-Backed RevEng.AI Raises $15M to Hunt Hidden Threats in AI-Generated Software

What RevEng.AI Actually Does

RevEng.AI builds a platform designed to analyze compiled software at the binary level — executables, firmware, and third-party binaries — and determine what is really inside them without requiring access to the original source code .

The company's core technology is BinNet, a proprietary AI model described as the world's largest for understanding binary machine code semantics . The model is trained to reverse-engineer compiled code and detect hidden vulnerabilities, backdoors, and malicious functionality — including threats that may have been inadvertently introduced by large language models generating unsafe code .

In practice, the platform acts as a verification layer for the software supply chain. When a defense contractor receives a firmware update from a supplier, or a bank deploys a containerized application built partly with AI-generated code, RevEng.AI's system can inspect the resulting binary and flag anomalies that static analysis or source-code audits would miss .

The startup's customer base already spans financial services, defense, national security, and critical infrastructure organizations, according to company statements cited in coverage of the round .

Why a $15 Million Series A Matters Now

The NATO-led investor syndicate is the signal

The composition of this round is unusually concentrated among defense- and intelligence-aligned investors. The NATO Innovation Fund led the round, marking one of the fund's more visible bets on a pure-play cybersecurity startup . In-Q-Tel, the not-for-profit venture arm that backs technologies relevant to the CIA and broader U.S. intelligence community, also joined — a repeat investor from the seed round .

When a 16-person London startup attracts simultaneous backing from NATO's venture fund and the U.S. intelligence community's investment vehicle, the signal is unambiguous: binary-level software verification is now considered a national-security capability .

The AI-generated code panic is real

The funding announcement explicitly frames the company's mission around the risks of AI-generated software . As large language models are increasingly used to write production code, organizations are losing visibility into what their software actually contains. RevEng.AI positions BinNet as a "binary-native verification layer" for this new reality — a claim the investor syndicate appears to have validated with capital .

The SiliconANGLE report on the round drew a direct comparison to Anthropic's own Mythos model, designed to identify cybersecurity risks in binaries, but noted that RevEng.AI has operationalized the concept into a deployable platform for enterprises and governments .

The speed from seed to Series A signals traction

RevEng.AI raised its $4.15 million seed round in June 2025, led by Sands Capital with support from In-Q-Tel Capital, IQ Capital, and Episode 1 . To close a Series A nearly four times larger just 11 months later — and to upgrade the lead investor to the NATO Innovation Fund — suggests that the company hit milestones that resonated with defense and enterprise buyers .

The capital will be directed toward expanding the engineering team, scaling the platform, and entering the U.S. market, where defense and critical infrastructure customers represent the most immediate commercial opportunity .

The Bigger Picture

Software supply chain attacks have become a boardroom-level concern since the SolarWinds breach, but the rise of AI-generated code introduces a subtler class of risk: code that is not intentionally malicious but structurally unsafe, generated by models that cannot reason about security properties in compiled output. RevEng.AI's bet is that binary-level AI analysis will become a standard layer in the software development lifecycle — and the $15 million Series A suggests that some of the world's most security-conscious institutions agree .