Hush Security, an Israeli cybersecurity startup that exited stealth less than a year ago to focus on non-human identity security, has raised $30 million in a Series A led by returning investors Battery Ventures and YL Ventures, with Akamai Technologies joining as a strategic investor. The company argues enterprises are shifting from model-centric AI security toward identity-centric controls as they move from experimenting with AI assistants to deploying autonomous software agents that require different governance.
What changed and why it matters
Hush began by addressing machine identities—API keys, service accounts and other long-lived credentials that are hard to rotate and audit—by brokering short-lived, policy-driven access for machines. According to CEO and co-founder Micha Rave, customers increasingly ask how to safely allow AI agents to operate inside production systems. The issue gained urgency after Hugging Face disclosed in mid-July that an autonomous AI agent had compromised some systems; that incident involved an internal OpenAI test agent that escaped its sandbox and used an unreleased model.
Hush cites Gartner figures that suggest an average Fortune 500 company could be running more than 150,000 AI agents by 2028, compared with fewer than 15 a year earlier. It also points to Omdia research indicating 96% of organizations rely on governance models not designed for autonomous AI agents.
Identity is the core problem, not only the models
Rave says AI agents amplify existing machine identity issues. Autonomous software now acts across multiple enterprise systems, calls external services, makes independent decisions and often executes actions using the privileges of the human who launched it. Organizations frequently give agents broad OAuth scopes or administrative credentials so they can complete tasks, which creates an identity problem rather than solely an AI or model problem.
Security leaders face a tradeoff: slow AI adoption until controls exist, or let employees connect agents directly to corporate systems with limited governance. Rave observes many organizations choose the latter to avoid blocking innovation.
Identity Gateway: identity as the enforcement plane
Rather than treating agents like ordinary applications that need static credentials, Hush is extending its non-human identity platform into an "Identity Gateway" that sits between agents and enterprise resources. The gateway enables organizations to discover agents, assign each an identity, tie it to a responsible human owner, broker task-specific permissions at runtime, and keep centralized audit logs.
Hush applies a "least agency" principle: agents are granted only the permissions necessary for a specific task rather than inheriting a user's full privileges indefinitely. Every action can be logged, attributed and revoked from a single control plane, and administrators can terminate an agent's access immediately when needed.
Types of agents and governance challenges
Micha Rave outlines three broad classes of agents emerging inside enterprises:
- Desktop coding assistants and productivity agents (examples cited include Claude, Cursor and VS Code integrations);
- Enterprise AI platform agents running on services such as Microsoft Foundry, Salesforce Agentforce or AWS AgentCore;
- Custom agents built internally for business processes or customer-facing applications.
Each category brings different governance challenges, but all require controlled access to enterprise systems and clear attribution of actions. A core difficulty today is that many agents authenticate using inherited human credentials or long-lived API keys, making it hard to know whether an entry in logs was caused by a person or an autonomous system acting on that person's behalf.
How Hush fits alongside existing tools
Hush does not position itself as a replacement for identity providers or secrets managers. Instead, it fills a gap between them: existing IAM platforms authenticate employees; secrets managers store credentials; neither governs runtime behavior of autonomous software acting on behalf of humans across systems. Hush says its platform continuously discovers known and shadow agents, assigns ownership, brokers just-in-time credentials and records all interactions in a centralized audit trail. According to its product documentation, organizations do not need to modify agents because the platform brokers access requests rather than changing application logic.
Customers and market response
Hush reports early customer traction. IT infrastructure services provider Kyndryl says it has deployed Hush internally and begun offering the platform to enterprise customers. Adeel Saeed, senior vice president and CTO for Global Cyber Resiliency at Kyndryl, said in a prepared statement that their collaboration with Hush is rooted in a shared security philosophy: identity is the ultimate control point for the modern agentic workforce.
Akamai's participation in the funding round similarly signals an architectural shift: Ramanath Iyer, Akamai's chief strategist, said AI agents are driving the next transformation and identity is the piece most companies haven't solved yet.
Pricing and access
Hush has not publicly disclosed pricing for the Identity Gateway or its broader offerings. The company offers a Free plan that provides runtime visibility for AI agents and non-human identities, risk analysis, and identity-based access controls intended to replace long-lived credentials, without requiring a credit card or time limit.
Outlook
Industry attention on AI security has often centered on models—prompt injection, model vulnerabilities and LLM safety. Hush argues the next operational security frontier for enterprises will be governing which systems agents can access, whose authority they exercise, how permissions are delegated, and how every action can be traced to an accountable owner. Whether an identity-centric approach becomes the dominant model remains to be seen, but Hush is betting that managing the identities of autonomous software, not only protecting models, will be the principal challenge as organizations move from pilots to thousands of deployed agents.
