Table of Contents
🌏 中文版
Funding Details
| Field | Value |
|---|---|
| Company | HiddenLayer (Austin, Texas) |
| Round | Series B |
| Amount | $100M |
| Lead investor | Delta-v Capital |
| Follow-on | Ten Eleven Ventures, Morgan Stanley, M12 (Microsoft's Venture Fund), Booz Allen Ventures |
| Valuation | Not disclosed |
| Total raised | At least $150M (including a $50M Series A in 2023) |
| Founded | Not disclosed (Series A closed in 2023) |
| Headcount | Not disclosed |
What This Company Does
HiddenLayer secures AI — protecting enterprise models, agents, and workflows from adversarial attacks, supply-chain poisoning, and malicious code injection.
Its platform spans four parts of the AI lifecycle: asset discovery (finding what models and agents are actually running inside an organization), AI supply chain security (scanning roughly 50 AI file formats to verify open-source models haven't been swapped for malicious lookalikes), attack simulation (red teaming), and runtime protection (real-time interception of prompt injection, agent manipulation, and tool misuse). This round's new product, Agent Harness Security, extends the Runtime Security module to protect autonomous coding agents that write, review, and ship code on their own.
Customers concentrate in financial services and large tech companies, with contracts spanning the US Department of Defense and intelligence community. One customer is described as a "leading frontier model provider" with "more than 700 million weekly users" — the CEO's hints point toward OpenAI or Anthropic. ARR grew more than 10x over the past year, reaching "tens of millions" of dollars, with over 90% of that growth coming from new customers.
What This Funding Signals
Implications for the Agent Ecosystem
The round funds three things: Agentic Runtime Security (giving enterprises visibility into what their agents actually do in production, flagging manipulation and unauthorized actions in real time), securing autonomous coding agents (which carry far more unsupervised authority than traditional dev tools, and where mistakes cost more), and expansion into Europe and EMEA. Gartner estimates enterprises will spend $2.83B this year on AI security products — 83% more than 2025 — climbing toward $4.78B next year. HiddenLayer's round is one bet on catching that budget shift.
What Investors Are Betting On
Delta-v Capital manages over $1.6B in assets and focuses on growth-stage infrastructure software and cloud services — a classic "the market has arrived" play. Three years ago, when HiddenLayer raised its first round, real-world examples of AI attacks at scale were hard to point to. Now that agents run in production en masse, the security question has shifted from "whether" to "how fast." The same week saw Noma and Zenity each raise over $100M, and Air Security close $50M on Sept. 1 — capital is stacking up fast in the agent-security niche, while large security vendors like Cisco, Palo Alto Networks, and Check Point have historically preferred to acquire rather than build (Robust Intelligence, Protect AI, Lakera) — suggesting this cohort's likely exit path is acquisition, not IPO.
Numbers Worth Watching
- ARR grew more than 10x in a year, over 90% from new customers — a sign of genuinely new demand emerging, not existing customers expanding usage.
- 39 granted patents and 65 pending — the research depth here is a real part of this valuation story, not just tooling but a first-mover threat taxonomy.
- 96% of organizations consider AI critical to core operations, yet nearly a third can't say for certain whether they've experienced an AI-related breach (per HiddenLayer's own 2026 AI Threat Landscape Report) — that gap is exactly what HiddenLayer sells against.
Watchlist Status
HiddenLayer is not yet tracked in the watchlist. Recommend adding it to section B7 (Agent Security / Governance), alongside already-tracked Zenity, Onyx Security, and Air Security, with tracking focus on: AI supply chain security + agent runtime protection + coding-agent security, $100M Series B led by Delta-v Capital.
Today's Takeaway
Three years ago, the biggest question around HiddenLayer's first round was whether AI attacks existed at meaningful scale at all. Now the question is whether security vendors can keep pace with how fast agent autonomy is expanding — coding agents already write, review, and ship their own code in production, and the traditional dev-tool security model simply doesn't cover that. The problem definition itself got rewritten inside six months.
References
Loading...