Fiddler AI Raises $30M Series C to Build Control Plane for Agent Governance
Fiddler AI closed a $30M Series C led by RPS Ventures, bringing total funding to $100M as enterprises confront the reality that AI agents require fundamentally different infrastructure than traditional software—not just monitoring, but active control.
The governance gap has real consequences. Agent failures can trigger regulatory fines, legal exposure, and brand damage, driving rapid adoption among Fortune 500 companies where AI oversight isn’t optional. Fiddler achieved 4x revenue growth over 18 months and earned AWS Pattern Partners status as organizations realize that deploying autonomous agents without proper controls creates existential risk.
The Agent Control Crisis
Traditional software monitoring assumes deterministic behavior: track metrics, catch exceptions, roll back deployments. AI agents break this model entirely. They make probabilistic decisions, use external tools, maintain state across sessions, and operate in feedback loops where small changes cascade into system-wide effects.
“The explosion of AI tools has created a significant gap in enterprise governance,” said Timothy Murphy, Partner at RPS Ventures. “Companies are deploying agents that interact with customers and make consequential decisions, but they’re monitoring them with fragmented point solutions built for simpler, deterministic use cases.”
The trust tax is immediate. Organizations scramble to implement manual reviews, human-in-the-loop approvals, and ad hoc evaluations before every release. Risk controls appear piecemeal through logging pipelines, policy checks, and escalation paths. The MIT research showing 95% GenAI pilot failures isn’t just about capabilities—it’s about the inability to govern autonomous systems at scale.
Four Pillars of Agent Control
Fiddler’s control plane addresses what founder Krishna Gade calls “distributed systems wearing a prompt.” Unlike traditional applications, a single agent request can fan out into retrieval calls, tool invocations, model calls across providers, and state updates that persist across sessions. Each step represents a potential failure point where plausible responses mask policy violations.
The platform provides four forms of control:
Causal Control: Complete reconstruction of decision chains through standardized telemetry and OpenTelemetry integration. Every agent action becomes traceable through spans that capture not just outputs, but the reasoning process that produced them.
Policy Control: Runtime enforcement rather than post-hoc monitoring. The system blocks unsafe tool calls, redacts sensitive outputs, enforces verification requirements, and routes high-risk actions through human oversight—all within the execution loop.
Measurement Control: Enterprise-grade evaluation that goes beyond “LLM-as-a-judge” approaches. Custom trust models deliver accuracy, safety, and compliance metrics calibrated for specific domains and continuously monitored for drift.
Economic Control: Financial guardrails for unbounded agent behavior. Unlike single model calls, agents can trigger cascading retries, tool usage, and secondary model invocations. Without governance, cost becomes a constraint rather than a choice.
Enterprise Validation at Scale
Regulated industries are driving adoption where compliance isn’t negotiable. Customers include Nielsen, Mastercard, US Navy, Ally Financial, and American Family Insurance—organizations where agent failures carry regulatory and reputational consequences.
“Fiddler has delivered unified observability, protection, and governance across agents and predictive models making it fundamental to our AI strategy,” said Karthik Rao, CEO of Nielsen. The platform achieved #1 ranking in AI Agent Security & Risk Management by CB Insights, validating its position as infrastructure rather than tooling.
The company’s growth reflects enterprise urgency around agent governance. Revenue increased 4x in 18 months as organizations moved from pilot projects to production deployments. AWS Pattern Partners status indicates deep integration across enterprise cloud infrastructure, positioning Fiddler as foundational rather than supplemental technology.
From Observability to Authority
Fiddler’s evolution mirrors the maturation of agent infrastructure. The company started with AI explainability, moved to data drift observability, and now provides active control. This progression reflects a fundamental insight: visibility without authority creates compliance theater.
“Observability gives you eyes, not authority,” explains Gade, drawing from his distributed systems background. Traditional monitoring tells you what happened; control planes enforce what should be allowed to happen. As agents become autonomous actors making business decisions, the distinction becomes critical.
The regulatory environment reinforces this direction. NIST’s AI Risk Management Framework, ISO/IEC 42001, and the EU AI Act all emphasize continuous governance over point-in-time audits. Compliance frameworks require record-keeping, traceability, and post-market monitoring—exactly what a control plane provides through a single system of record.
Infrastructure Standardization Emerges
The funding accelerates Fiddler’s vision to become the neutral control plane for compound AI systems. Rather than vendor-specific solutions, the platform provides infrastructure-agnostic governance across frameworks and model providers.
This standardization matters because agent architectures are converging on common patterns: policy-driven loops, tool orchestration, multi-agent collaboration, and memory systems. Despite implementation differences, all require the same governance primitives: telemetry capture, policy enforcement, evaluation pipelines, and audit trails.
“Every major platform shift requires new infrastructure,” said Gade. “When companies moved to the cloud, they needed orchestration layers. As they deploy autonomous agents, they need a control plane.”
The market timing aligns with enterprise deployment cycles. Organizations are moving from proof-of-concepts to production systems where governance becomes mandatory, not optional. Fiddler’s control plane provides the infrastructure layer that enables this transition.
The control plane category emergence signals infrastructure maturity for autonomous AI. Just as Kubernetes emerged when container orchestration became necessary, control planes become inevitable as organizations deploy fleets of decision-making agents. Fiddler’s funding positions it to define this foundational layer, providing the governance infrastructure that transforms agent pilots into production systems.
For teams building agent orchestration platforms, Overclock provides complementary workflow automation that integrates with governance infrastructure, enabling reliable execution while maintaining enterprise oversight and control.