Companies are deploying AI agents faster than they can track what those agents are doing. JetStream Security just raised $34 million to solve that problem before it becomes a disaster.
The seed round, led by Redpoint Ventures with participation from CrowdStrike’s Falcon Fund, closed within weeks - a sign of how urgent investors see the AI governance gap. Angel backers include CrowdStrike CEO George Kurtz, Wiz CEO Assaf Rappaport, and Okta co-founder Frederic Kerrest. When the people who built enterprise security infrastructure put their money into AI governance, the signal is clear.
The Problem: Nobody Knows What’s Happening
Ask most enterprise security teams these questions about their AI deployments:
- Which data did that agent access?
- Which model made that decision?
- Who approved the workflow?
- How much did that run cost?
They can’t answer. JetStream’s early customer work suggests the visibility gap is worse than companies realize.
Around 70% of organizations believe they have “shadow AI” - unauthorized tools employees use without IT approval. JetStream’s founders say the actual number is typically higher. Employees sign up for AI services, connect them to company data, and build workflows without going through security review. By the time IT notices, sensitive data has already flowed through systems nobody controls.
The Founders Know This Space
JetStream’s team isn’t new to enterprise security. CEO Raj Rajamani, COO Jared Phipps, CTO Jatheen “AJ” Anand, and Chief Architect Venu Vissamsetty come from CrowdStrike, SentinelOne, Cohesity, McAfee, Attivo Networks, and Cylance. They’ve shipped products through IPOs at major security companies.
That experience matters because AI governance isn’t just a technical problem - it requires understanding how enterprises actually buy and deploy security tools. JetStream is already working with Fortune 500 customers, suggesting the product resonates with security buyers who’ve seen this movie before with cloud computing.
How AI Blueprints Works
JetStream’s approach maps AI activity in real time using what the company calls “AI Blueprints” - dynamic graphs showing relationships between agents, models, data sources, tools, and the people responsible for each.
The system flags activity that drifts from approved purposes, tracks costs at the agent level, and provides runtime visibility rather than static architecture diagrams. When an AI agent starts accessing data it shouldn’t, or costs balloon on a workflow nobody authorized, the platform surfaces it.
The distinction from traditional monitoring matters. Conventional security tools watch for threats. AI governance tools need to watch for drift - authorized systems doing unauthorized things, or costs scaling beyond what anyone approved.
Why This Matters Now
Global AI spending is projected to hit $650 billion this year, according to Fortune. But 93% of executives report challenges implementing AI governance and security controls. That disconnect - massive spending with minimal visibility - is why investors see AI governance as a category about to explode.
The pattern mirrors what happened with cloud security a decade ago. Companies moved workloads to AWS and Azure faster than their security tools could follow. A generation of startups (including CrowdStrike and Wiz, whose leaders now back JetStream) emerged to fill the gap. AI is following the same trajectory, just faster.
The Stakes
Half of CEOs acknowledge their roles could be at risk if AI investments underperform. Yet most can’t answer basic questions about what their AI systems are doing. That’s a governance crisis waiting to happen - whether through data breaches, regulatory violations, or runaway costs.
JetStream is betting that enterprises will pay for visibility before regulators force them to. Given the pace of AI deployment, they may not have much time either way.