Nick Warner has a knack for timing. The former SentinelOne COO who helped steer that cybersecurity darling through its 2021 public offering knows when the market's ready for a new category—and when investors are ready to write big checks before anyone's proven it exists.
On July 20, Warner's newest venture, Neo Security, surfaced from stealth with $100 million in funding. The thesis? Enterprises have no idea what AI agents are running inside their networks, how they're configured, or what they're doing. Warner calls them "zombie agents," and he's convinced the Boston-based startup can tackle what may become one of the messier problems in AI adoption: visibility and control.
Andreessen Horowitz led the seed. Bessemer Venture Partners led the Series A, though the company hasn't disclosed how the $100 million breaks down between the two rounds. Craft Ventures and Merlin Ventures joined the round. It's the kind of compressed fundraising structure that's become standard in enterprise security, where credible teams can secure growth capital well before they've hit the milestones that once defined product-market fit.
Whether Neo can deliver on the promise is another question entirely.
The SentinelOne Alumni Network Strikes Again
Warner didn't build Neo alone. He brought along Shlomi Salem and Eran Shirazi, both of whom carry the kind of technical credentials that make venture partners lean forward in pitch meetings.
Salem spent more than a decade at SentinelOne running detection engineering and threat research—the unglamorous, critical work of building threat actor profiles that enable autonomous response. Shirazi's resume is equally sharp: he co-founded EasySend and, before that, led a vulnerability research group inside Unit 8200, Israel's storied intelligence corps that's spawned a disproportionate number of cybersecurity unicorns.
Fortune reported that a16z general partner Zane Lackey and Bessemer's Elliott Robinson took board seats, a detail conspicuously missing from Neo's official press release. (The company did not respond to questions about board composition.)
Agentic Software Control—or Security Theater?
Neo describes what it does as "Agentic Software Control," which is either a genuinely new category or a clever rebrand of endpoint monitoring, depending on your tolerance for buzzwords.
The pitch centers on real-time visibility into AI agents, AI-enabled applications, browser extensions, and legacy software that's picking up agentic capabilities. Neo's approach: deploy a lightweight endpoint sensor—25 megabytes, no kernel module, no reboot required—that promises full inventory scanning in under 15 minutes.
Once installed, the sensor tracks what agents exist, how they're configured (permissions, credentials, scope of data access), and what happens when they execute. The system purports to distinguish human-initiated actions from autonomous agent behavior, then enforce policy at the endpoint before any data or system access occurs.
The enforcement piece relies on an LLM-powered policy authoring system that generates guardrails from observed behavior. Backing it up is "Neoverse," Neo's term for a knowledge base cataloging agentic software capabilities, risks, and version-specific behaviors. The company emphasizes on-device analysis—a privacy angle that may appeal to regulated industries skittish about shipping behavioral telemetry to external platforms.
It's an elegant architecture on paper. How it performs in the chaos of actual enterprise environments is harder to say. Neo hasn't released technical documentation, third-party testing results, or benchmark data comparing its approach to emerging alternatives. For now, prospective buyers are betting on the founders' track record.
Early Customers, Few Details
Warner told Fortune that Neo's already running in financial services, healthcare, insurance, and manufacturing organizations. He didn't name customers, share revenue figures, or disclose deployment scale—standard opacity for a company barely out of stealth. eWeek separately reported pilot activity in financial services, transportation, and energy sectors, suggesting Neo's targeting critical infrastructure where uncontrolled AI agents could trigger compliance headaches or operational disasters.
The company's careers page listed 37 open roles in late July, split between Boston and Tel Aviv, spanning sales, engineering, legal, and operations. Warner mentioned to the Boston Business Journal that he plans to hire north of 20 people by year's end—a measured pace that suggests Neo's not in full growth-stage sprint mode yet.
A Crowded, Chaotic Market
Neo's $100 million debut landed in the middle of a remarkable funding frenzy. Between mid-June and late July, at least five companies in adjacent categories pulled in substantial rounds: NewCore secured $66 million for agent identity and governance on June 15. Valarian raised $50 million for sovereign infrastructure control on July 14. Oak closed $60 million for an identity OS spanning humans, AI, and machines on July 16. And Onyx Security wrapped a $113 million Series B on July 30, hitting an estimated $640 million valuation, according to Calcalist.
The timing makes sense, even if it feels frothy. Industry projections suggest that a significant portion of enterprise applications will embed AI agents by the end of 2026, up from a small fraction in 2025. NIST launched an AI Agent Standards Initiative in February, focusing on security and identity research—a signal that regulators are finally waking up to governance frameworks that didn't exist 18 months ago.
What sets Neo apart, at least in theory, is the endpoint focus. That's a direct inheritance from Warner and Salem's SentinelOne tenure, where autonomous threat detection depended on granular visibility into device-level activity. Whether that model translates to distributed, cloud-native environments—where many agentic workflows actually execute—remains an open question. And probably the question that determines whether Neo's $100 million bet was prescient or premature.
What Comes Next
Neo says it's using the capital to scale go-to-market and engineering rapidly. The company held closed-door briefings and a launch event at Black Hat USA in Las Vegas in early August, courting the CISO crowd now staring down agent sprawl without mature tooling.
For now, Neo's running on founder credibility and a thesis that resonates: enterprises need visibility and control before agent proliferation spirals beyond their ability to govern it. That's a compelling story. Whether it's a $100 million story will depend on how well the technology actually works—and how quickly Neo can prove it in production environments where security theater gets exposed fast.
The zombie agents are multiplying. Warner's betting he can see them before anyone else does.
