#

Key takeaways

• HiddenLayer announced a $100 million Series B.

• Delta-v Capital led the round.

• The company plans to deepen its enterprise platform.

• The platform includes Agentic Runtime Security and Agent Harness Security.

• The source does not establish Morocco-specific availability or partnerships.

HiddenLayer raises $100M in Series B

HiddenLayer announced a $100 million Series B on September 2, 2026. Delta-v Capital led the round. Ten Eleven Ventures, Morgan Stanley, M12, Microsoft's Venture Fund, and Booz Allen Ventures also participated.

The company said it will use the funding to deepen its enterprise platform. The announcement highlights two named capabilities: Agentic Runtime Security and Agent Harness Security. These are described as protections for AI coding agents at runtime.

The source frames the funding as a signal for enterprise AI security. It points to a specific operational concern. Teams should think beyond model choice and consider runtime security as part of adoption.

What the announcement says

The input data focuses on platform expansion. It does not provide product pricing, deployment details, or technical specifications. It also does not describe customer outcomes or performance claims.

The announcement is narrow in scope. It tells readers that HiddenLayer is investing in security capabilities for AI agents. It does not say how those capabilities work in practice.

That matters for readers evaluating AI agents. Security decisions often happen after model selection. This source suggests that runtime protection should be part of the same conversation.

Why runtime security matters here

The most specific detail in the source is runtime protection for AI coding agents. That means the security focus is not only on the model itself. It also covers how the agent behaves while it is running.

This is an operational consideration, not a broad market claim. The source does not say that all AI systems need the same controls. It only shows that HiddenLayer is building for this use case.

For decision-makers, the practical lesson is simple. If an AI agent will act during live workflows, security should be evaluated at that stage too. That is the clearest implication supported by the input.

Governance and operational considerations

The announcement implies a need for layered review. Model selection alone may not be enough when agents can execute tasks at runtime. Security teams may need to assess agent behavior, not just model output.

The source does not provide a governance framework. It does not mention audits, compliance, or policy controls. So any broader governance approach should be treated as an assumption, not a fact from the announcement.

Still, the funding round shows where vendor attention is going. Enterprise buyers may want to ask whether runtime protections are included in their AI stack. They may also want to confirm how those protections apply to coding agents.

Morocco relevance

The source reports no Morocco-specific product availability, partnership, or market detail. The conditional lesson is global: if a team adopts AI agents, runtime security should be reviewed alongside model choice.

What is not established by the source

The input does not say whether HiddenLayer's platform is available in any specific country. It does not name customers, pilots, or regional partners. It also does not establish any Morocco-specific use case.

That limits how far the announcement can be interpreted. The safest reading is that HiddenLayer is expanding its security platform for enterprise AI agents. The broader lesson is about timing and scope of security planning.

Bottom line

HiddenLayer's $100 million Series B is a clear sign of continued investment in AI security. The company is focusing on runtime protection for AI coding agents. That makes the announcement relevant to any organization evaluating agentic systems.

The source supports one main takeaway. AI adoption should include security at the point of execution, not only at the point of selection. Everything beyond that should be treated as an assumption unless further details are provided.