The convergence of tech giants around a single investment thesis doesn't happen often. So when Microsoft's venture arm and IBM Ventures both wrote checks to HiddenLayer last September—a $50 million Series A that the Austin startup claims represented the year's largest for an AI security company—it signaled something about where the enterprise AI market was heading.
Or perhaps more accurately, where its anxieties were heading.
Chris "Tito" Sestito, HiddenLayer's CEO, has been watching adversarial AI attacks evolve since his days at Cylance, the endpoint security firm later acquired by BlackBerry. Back then, machine learning models were mainly academic curiosities or confined to research labs. Now they're embedded in production systems at banks, defense contractors, and cloud platforms—and they're vulnerable in ways that traditional cybersecurity tools weren't built to address.
"The attack surface has fundamentally changed," Sestito noted around the time of the funding. The company's pitch centers on defending AI systems from threats that can poison training data, manipulate predictions, or extract proprietary models entirely. It's a category that barely existed five years ago.
A Roster That Reads Like an AI Deployment Map
The September 2023 round, co-led by M12 (Microsoft's Venture Fund) and Moore Strategic Ventures, pulled in an eclectic mix of corporate investors: Booz Allen Ventures, IBM Ventures, Capital One Ventures, and returning backer Ten Eleven Ventures. The lineup reads less like a typical VC syndicate and more like a cross-section of industries scrambling to operationalize AI while managing new risks.
Capital One, for instance, has been publicly vocal about deploying machine learning for fraud detection and credit decisioning. Booz Allen, meanwhile, framed its investment as a strategic play tied to federal government AI initiatives—no small consideration given HiddenLayer's contracts with the U.S. Air Force and Space Force.
IBM's involvement arrived a few months before the company announced a $500 million Enterprise AI Venture Fund in November 2023, suggesting the tech incumbent was already mapping out the AI security landscape. Microsoft's M12, naturally, has its own motivations: Azure customers are building AI applications at scale, and those workloads need protection.
The Series A brought HiddenLayer's total disclosed capital to $56 million, building on a $6 million seed round from July 2022.
The Technical Challenge (and Sales Pitch)
HiddenLayer's platform is built around three core products: Machine Learning Detection & Response for runtime monitoring, Model Scanner for static analysis, and AI Attack Simulation for automated red teaming. The company's differentiator, according to its marketing materials, is a turnkey approach that doesn't require direct access to raw data or proprietary models—an important consideration for enterprises skittish about exposing their AI crown jewels to third-party scrutiny.
Whether that architecture holds up under real-world adversarial conditions remains an open question. The field is young enough that peer-reviewed benchmarks are still emerging, and what works against today's attack techniques may not scale against tomorrow's.
Still, the company had secured partnerships with Intel and Databricks by late 2023, and claimed Fortune 100 customers in finance and government. It also picked up industry accolades: RSA Conference's Innovation Sandbox "Most Innovative Startup" award in April 2023, followed by Gartner Cool Vendor recognition for AI Security in 2024 and inclusion in the 2025 Gartner Market Guide for AI TRiSM—the acronym-heavy shorthand for Trust, Risk, and Security Management in AI systems.
Growing Into a Category Still Being Defined
At the time of the announcement, HiddenLayer planned to scale from roughly 50 employees to 90 by year-end 2023, using the capital to build out both engineering teams and go-to-market operations. The MLSec Platform—HiddenLayer's umbrella term for its product suite—was slated for continued development, though specifics on roadmap priorities were thin.
The competitive landscape is still forming. Startups like Robust Intelligence, CalypsoAI, and Troj.ai occupy adjacent spaces, each pitching slightly different approaches to AI model security. The broader question is whether this becomes a standalone category or gets absorbed into existing security platforms as vendors like CrowdStrike, Palo Alto Networks, and others expand their portfolios.
For now, enterprises are hedging. The shift from experimental AI projects to production deployments has created urgency around governance and security—areas where retrofitting traditional tools often falls short. HiddenLayer's bet, backed by some of the biggest names in enterprise tech, is that organizations will pay for purpose-built defenses.
Whether that thesis plays out depends less on the elegance of the technology and more on how quickly AI failures become headline risks. Given the rate at which generative AI has already triggered regulatory scrutiny and public skepticism, the market may be arriving faster than anyone expected.
