Cymphony Emerges With $30M Series A To Police Enterprise AI Agents

Sequoia and SMBC lead a $30M Series A for Cymphony, a stealth-exit startup building an identity layer that gives enterprise security teams control over AI agents accessing sensitive data.

Cymphony Emerges With $30M Series A To Police Enterprise AI Agents

New York and Tel Aviv-based Cymphony emerged from stealth on September 9, 2026 with a $30 million financing round led by Sequoia Capital and SMBC Fin Atlas Beyond Fund, valuing the enterprise AI-agent security startup at more than $100 million post-money.

The round

The Series A includes a $25 million lead check from Sequoia and SMBC alongside a previously undisclosed Sequoia seed investment, for a total of $30 million in disclosed funding. Founders Shy Dekel (CEO), Idan Berkovits and Edi Gotlieb — veterans of Israel's cybersecurity ecosystem — will use the capital to expand engineering and go-to-market ahead of what the company describes as an accelerating enterprise deployment cycle.

Why AI agents need their own security stack

Cymphony argues that identity and access management tools built for humans are wrong-sized for the current wave of AI agents that read email, act on databases and negotiate with SaaS APIs at machine speed. The company's core product is a 'workforce graph' that inventories every non-human identity in an enterprise, maps what data it can see, and enforces guardrails on what actions it can take. Early customers named on the record include KKR, Syngenta, Cass Information Systems and Athennian.

Cymphony workforce graph dashboard for AI agent identities

A hot but crowded market

Cymphony joins a wave of new companies — including some featured in Bynario's recent pre-seed and JetStream's clearance product — racing to build a security stack purpose-built for agentic AI. Sequoia's move signals that at least one top-tier fund sees this as a foundational layer rather than a feature.

What comes next

Cymphony plans to launch a self-service tier for mid-market customers, expand integrations into Microsoft, Google and Anthropic agent runtimes, and hire aggressively in New York and Tel Aviv. Dekel says the target is a full runtime authority layer that treats every AI agent as a first-class identity with just-in-time permissions.

Reporting based on coverage from TechCrunch, Reuters and PR Newswire.

Category: Funding & Investments

Tags: venture capital funding Cybersecurity Enterprise AI AI Agents agentic AI AI safety

Related Articles