Onyx Security, an Israeli cybersecurity startup founded in 2024 by elite alumni of the IDF’s Unit 8200 and Nvidia AI research teams, has secured $35 million in a Series A funding round led by AI-native venture capital firm Conviction to develop its groundbreaking platform that provides comprehensive security, governance, and operational control over autonomous AI agents increasingly embedded across enterprise software development pipelines, IT automation workflows, customer service interactions, and mission-critical business processes where these dynamic decision-making systems introduce unprecedented attack surfaces and compliance risks that traditional cybersecurity architectures fundamentally fail to address.

Series A Funding and Investment Syndicate

The $35 million Series A investment, which elevates Onyx Security’s total capital raised to $40 million following a $5 million seed round from Cyberstarts in 2024, arrived at a pivotal moment when enterprises worldwide confront the chaotic proliferation of AI agents operating with varying degrees of autonomy across fragmented technology stacks, attracting participation from prominent angels spanning cybersecurity veterans and AI research luminaries who recognize the startup’s potential to establish category leadership in governing what Conviction partner Sarah Guo describes as “systems that make decisions, access sensitive information, and integrate into critical business processes” rather than merely protecting static software tools ill-equipped for real-time behavioral analysis and contextual risk assessment.[web:query] Conviction—founded by Guo after her tenure as General Partner at Greylock—brings specialized conviction in AI infrastructure plays, having backed foundational companies navigating enterprise adoption inflection points where technical excellence intersects with C-suite governance imperatives demanding measurable return quantification alongside ironclad security assurances amid 2026’s regulatory scrutiny over autonomous system accountability.

This funding infusion arrives precisely when AI agent deployments have transitioned from experimental pilots to production infrastructure at Fortune 500 scale, creating governance vacuums that Onyx addresses through continuous asset discovery, real-time policy enforcement, and human-in-the-loop escalation mechanisms positioning the Tel Aviv-headquartered firm to capture substantial market share in what Cyberstarts partner Hila Zigman characterizes as “an entirely new attack surface” requiring purpose-built solutions rather than retrofitted legacy tools struggling to parse dynamic agent behaviors across cloud-native environments.

Elite Founding Team with Unparalleled Credentials

Co-founded by Maxim Bar Kogan—a seasoned cybersecurity executive and Unit 8200 veteran who brings operational battle-hardening from Israel’s premier signals intelligence unit—and Gil Elbaz—a veteran AI researcher who reported directly to Nvidia’s CTO while contributing to IDF AI research initiatives—the leadership team combines practitioner insight from securing nation-state grade infrastructure with frontier model expertise essential for governing agentic systems that autonomously synthesize external data streams, execute API calls across enterprise boundaries, and generate code or content requiring continuous behavioral monitoring rather than static vulnerability scanning.[web:query] Bar Kogan’s insistence that “AI agents are not too independent” captures the platform’s dual mandate: preventing rogue actions through granular permission controls while maximizing business value through measurable productivity quantification, addressing CEO pain points where AI investments register prominently on capex ledgers yet deliver opaque returns invisible through traditional KPI frameworks ill-suited for assessing autonomous workflow contributions.

Currently employing over 70 engineers across Israel, the United States, and Canada, Onyx demonstrates sophisticated global talent acquisition strategy positioning North American go-to-market teams alongside Tel Aviv R&D core, with Bar Kogan’s planned relocation to the U.S. signaling acceleration toward enterprise sales motions targeting CISOs and AI governance officers confronting board-level mandates to demonstrate compliance alongside ROI across increasingly agent-saturated technology portfolios.

Revolutionary Platform Architecture and Guardian Agent

Central to Onyx Security’s value proposition resides the Onyx Guardian Agent—a sophisticated meta-agent continuously monitoring deployed AI systems through asset discovery mapping every autonomous actor across development pipelines, customer-facing interfaces, and internal automation workflows, enforcing granular usage policies governing data access patterns, API execution permissions, external integrations, and decision thresholds while detecting anomalous behaviors through behavioral baselines established via machine learning models trained on enterprise-specific agent interaction patterns.[web:query] Unlike perimeter-focused security tools struggling with ephemeral cloud-native agents lacking static endpoints, Onyx operates as persistent behavioral supervisor capable of blocking dangerous actions mid-execution, dynamically reducing permissions through just-in-time access controls, or routing high-risk decisions for human approval, creating governance flywheel where safe scaling amplifies business value while containing blast radius from compromised agents potentially exfiltrating proprietary datasets, generating malicious code, or executing unauthorized financial transactions.

The platform’s real-time policy enforcement engine distinguishes nascent agent governance category by continuously validating compliance across dynamic execution contexts—unlike scheduled vulnerability scans missing ephemeral threats—while providing executive dashboards quantifying ROI through metrics demonstrating accelerated code review cycles, reduced customer support resolution times, and IT automation throughput improvements directly attributable to safely governed agent deployments rather than siloed productivity experiments.

Critical Market Timing Amid AI Agent Proliferation

Onyx Security confronts enterprise reality where AI agents have metastasized across operational infrastructure: software engineering teams deploy autonomous code generation agents interfacing with private Git repositories containing proprietary algorithms and customer data; customer service organizations orchestrate multi-agent systems handling billing disputes, technical escalations, and compliance documentation across regulated verticals; IT operations automate incident response through agents accessing Active Directory, cloud entitlements, and configuration management databases—each introducing decision-making autonomy and external data synthesis capabilities transforming static software risks into dynamic behavioral threats demanding continuous governance rather than periodic vulnerability assessments.[web:query] Traditional cybersecurity paradigms collapse against these realities: endpoint detection struggles with serverless agent executions; data loss prevention tools miss encrypted outbound API calls; identity platforms cannot govern nonhuman principals operating across ephemeral cloud credentials—creating trillion-dollar risk exposure as enterprises race toward agentic architectures promising 10x engineering velocity absent corresponding control frameworks.

Bar Kogan articulates the C-suite calculus succinctly: “If I am the CEO of an organization today, I want to show progress in AI,” confronting reality where investments register visibly through vendor contracts and compute costs yet deliver returns obscured by fragmented deployment patterns across engineering, operations, and customer teams requiring unified governance dashboards demonstrating risk-adjusted productivity amplification rather than speculative transformation narratives.

Aggressive U.S. Expansion and Competitive Positioning

With dedicated U.S. sales organization already operational and Bar Kogan’s relocation imminent, Onyx targets massive greenfield opportunity coinciding with 2026 regulatory inflection points where SEC disclosure requirements, EU AI Act compliance deadlines, and NIST agent governance frameworks compel CISOs to demonstrate enterprise-wide visibility and control over autonomous systems beyond traditional software supply chain security.[web:query] Competitive landscape features adjacent players—Jazz securing AI data pipelines ($43M raised), Fig strengthening resilience infrastructure ($30M), Reclaim automating remediation ($20M)—yet Onyx carves distinct leadership governing agent behaviors rather than adjacent infrastructure layers, positioning as control plane for agent mesh architectures where enterprises orchestrate hundreds of specialized agents requiring centralized policy enforcement and anomaly detection across distributed execution environments.

Conviction’s Guo endorsement validates category creation potential: “From the very first moment, it was clear… that Maxim and Gil understand the depth of this change and are building a solution designed for the AI era,” distinguishing Onyx’s native architecture from incumbents retrofitting SIEM platforms or CASBs for agent governance—a doomed strategy against behavioral complexity demanding purpose-built behavioral analytics and real-time intervention capabilities.

Strategic Roadmap and Category Leadership Trajectory

$35 million deployment accelerates three-pronged expansion: North American enterprise sales targeting Global 2000 enterprises through CISO and AI governance officer motions; R&D investment scaling Guardian Agent capabilities toward multi-agent mesh governance, cross-cloud policy federation, and adversarial attack simulation; strategic partnerships embedding Onyx controls natively within agent frameworks like LangChain, LlamaIndex, and enterprise platforms from Microsoft, ServiceNow, and Salesforce where agent orchestration emerges as table stakes for 2027 competitive differentiation.[web:query] Near-term milestones include SOC 2 Type II attestation, FedRAMP authorization trajectory, and integrations with leading agent observability platforms establishing Onyx as default control layer for enterprise agent meshes while longer-term vision contemplates category-defining acquisition by strategic acquirers confronting existential agent governance gaps across hyperscale customer bases.

Onyx Security arrives not merely as cybersecurity vendor but foundational control infrastructure for agentic enterprise, governing trillion-dollar transformation from human-mediated workflows toward autonomous decision architectures where behavioral security determines competitive survival—a positioning validated by elite team credentials, conviction backing, and pristine timing capturing enterprises confronting boardroom mandates to demonstrate risk-adjusted AI ROI through unprecedented agent governance capabilities.