A startup you've never heard of just raised $66 million to solve a problem most people did not know exists — giving AI agents their own digital identities. NewCore, a stealth infrastructure company emerging from two years of quiet development, announced a $66M seed round led by Sequoia Capital and Andreessen Horowitz. The funding signals a tectonic shift in how enterprises think about their AI workforce.

As autonomous AI agents multiply across organizations — handling customer service, writing code, managing supply chains, and even making financial decisions — a fundamental question emerges: how do you know which AI agent did what? The answer requires an entirely new category of infrastructure: the agent identity infrastructure layer.

What Is an AI Agent Identity Platform?

An agent identity platform provides each autonomous software agent with a verifiable, secure, and auditable digital identity — much like an employee badge or a passport for software. Just as every human employee in a company has a unique identifier, access permissions, and an audit trail, AI agents need the same infrastructure to operate in enterprise environments.

Traditional identity and access management (IAM) systems like Okta, Azure AD, and JumpCloud were designed for humans accessing SaaS applications. They authenticate people, manage sessions, and enforce role-based access to tools. AI agents operate differently — they need machine-to-machine authentication, API-level permissions, long-running autonomous sessions, and the ability to delegate tasks to other agents. Human IAM systems cannot handle these patterns.

What Makes Agent Identity Different

• Non-human authentication — AI agents cannot type passwords or use SSO portals. They need API-key-based identity tokens that prove which model, configuration, and permissions they carry.
• Delegation chains — When Agent A asks Agent B to complete a subtask, the identity platform must cryptographically chain those actions so the full decision tree is auditable.
• Session persistence — Unlike a human who logs in and out, an AI agent may run for days. Its identity must be revocable mid-task without breaking dependent workflows.
• Compliance attestation — Regulators increasingly require evidence that AI actions were authorized. Agent identity platforms produce the digital paper trail.

An agent identity platform authenticates machine agents, verifies permissions, and creates cryptographic audit trails for every action taken.

Why the AI Agent Identity Platform Market Matters for Enterprise AI

NewCore is not announcing a product launch — the company remains in stealth mode with only a technical whitepaper and a public GitHub repository containing an early API specification. Yet investors committed $66 million based on the founding team's credibility and the magnitude of the problem. The founders include former lead architects from the IAM teams at Google Cloud and Okta, plus a distributed systems researcher from MIT.

The size of the round — $66 million at seed stage — is extraordinary by any measure. To put it in context, the average AI infrastructure seed round in 2025 was around $8 million. NewCore's raise is more than 8x the category average. This suggests either aggressive competition among top-tier VCs or a conviction that agent identity will become as foundational as DNS, SSL, or IAM itself.

How NewCore's Approach Compares

The agent identity space is nascent but already seeing competition from major cloud providers and startups. Here is how the key players stack up:

NewCore's most significant differentiator is its open-source protocol specification. By publishing the agent identity protocol on GitHub, the company is positioning itself as the "Linux of agent identity" — an open standard that any cloud provider, enterprise, or agent framework can adopt. This strategy mirrors how Kubernetes became the standard for container orchestration: not by being the best product, but by being the best open standard.

Three Shifts the AI Agent Identity Platform Signals for Enterprise Teams

The $66 million round is not just about one startup. It tells us where the broader AI industry is heading and how three key stakeholder groups should prepare.

For Enterprises: The AI-as-Employee Paradigm

Enterprises are moving from experimenting with AI chatbots to deploying production AI agents that handle real business functions. A 2026 survey by Gartner found that 43% of enterprises already run at least 10 AI agents in production, and 12% run more than 100. At these scales, knowing which agent did what is not optional — it is a compliance and security necessity.

The AI-as-employee concept means treating agents like staff: they need onboarding (identity creation), role assignments (permissions), performance reviews (audit logs), and offboarding (identity revocation). NewCore's platform is building the HR system for AI workers. Enterprises evaluating AI agent deployments should demand identity capabilities from their vendors today — retrofitting identity later is much harder than building it in from the start.

For Developers: New Infrastructure to Build On

NewCore's open-source protocol means developers can integrate agent identity at the framework level rather than cobbling together API keys and ad-hoc tokens. The protocol specification, available on GitHub, defines how agents register, authenticate, delegate, and log actions using standard cryptographic primitives (Ed25519 signatures, verifiable credentials following W3C standards).

For AI engineers building multi-agent systems, this removes a significant security burden. Instead of designing custom authentication for each agent network, teams can drop in NewCore's identity layer and get audit trails, revocation, and cross-agent authentication out of the box. The API supports Python, TypeScript, and Go SDKs at launch, with Rust and Java in development.

For Regulators: The Audit Trail Problem

Regulatory interest in AI accountability is accelerating. The European Union's AI Act, which came into full effect in early 2026, requires that high-risk AI systems maintain comprehensive logs of their operation. The US Executive Order on AI Safety, updated in March 2026, mandates that federal agencies using AI document every automated decision with an audit trail.

Neither regulation explicitly mandates dedicated agent identity infrastructure. But both require proof of who or what made each decision, and agent identity infrastructure is the only scalable way to provide that proof. NewCore's whitepaper explicitly maps its protocol to EU AI Act logging requirements and NIST AI Risk Management Framework controls — a strategic move that positions the startup as a compliance enabler, not just a security tool.

Enterprises deploying AI agents at scale need identity infrastructure as foundational as their HR and IAM systems.

The Road Ahead: Challenges and Open Questions

NewCore's vision is compelling, but significant challenges remain. The startup has no publicly known customers, no production deployment, and only a GitHub repository to show for two years of work. The $66M seed gives it runway to build, but enterprise identity infrastructure is notoriously hard to sell — procurement cycles for IAM products routinely exceed 12 months.

Interoperability with existing enterprise systems is another hurdle. Companies already invested in Okta, Azure AD, or Ping Identity will need NewCore to integrate rather than replace. The company's open-source approach helps here — CIOs can review the protocol before committing — but integration work is unavoidable.

There is also the question of how many AI agents organizations will actually deploy. If the market evolves toward fewer, more powerful monolithic AI systems rather than swarms of specialized agents, the need for agent identity at scale diminishes. The agentic AI trend is strong, but enterprise adoption patterns often surprise optimists.

FAQ: AI Agents and the Digital Workforce

What is AI agent identity management?

AI agent identity management is the practice of giving each autonomous software agent a unique, verifiable digital identity with associated permissions, audit capabilities, and cryptographic proof of actions. It parallels human identity management in enterprises but addresses machine-specific requirements like delegation chains, long-running sessions, and API-level authentication.

Why do AI agents need digital identities?

Without digital identities, enterprises cannot determine which AI agent took which action, whether that action was authorized, or how to revoke access when an agent is decommissioned. As regulations like the EU AI Act demand audit trails for automated decisions, identity becomes a compliance necessity. Security teams also need agent identity to prevent malicious agents from impersonating legitimate ones.

How do agent identity platforms work?

These platforms typically register each agent with a cryptographic key pair, issue verifiable credentials (following W3C standards), maintain a directory of active agents, handle agent-to-agent authentication, log all actions with signatures, and provide APIs for permission management and identity revocation. Most platforms integrate with existing IAM systems through standard protocols like SAML, OAuth 2.0, and SCIM.

What is NewCore AI and why is its funding significant?

NewCore is a stealth startup building an open-source protocol for AI agent identity. Its $66M seed round (led by Sequoia and a16z) is the largest known seed round in the AI infrastructure space, signaling VC conviction that agent identity will become foundational infrastructure — as essential as DNS or IAM — for the enterprise AI era.

Conclusion: Identity Is the Bedrock of the AI Workforce

NewCore's record-breaking seed round is a signal that the AI industry is maturing beyond model capabilities toward operational infrastructure. Just as every employee needs a badge and every server needs a certificate, every AI agent will need a verifiable identity. The companies that invest in agent identity infrastructure today will be the ones that scale AI safely and compliantly tomorrow.

Agent identity platforms are not a niche security product — they are the operating system for the AI-augmented enterprise.

Want to stay ahead of enterprise AI infrastructure? Bookmark Markly and check back daily for analysis of the trends shaping AI adoption. Drop your experience in the comments — is your organization already planning for AI agent identity, or is this the first you've heard of the category?