Empirik raises $21M to predict IT outages

Empirik, a new AI startup backed by Sequoia Capital’s incubation arm, has launched with $21 million in funding to predict IT outages before they happen. The company uses machine learning to analyze infrastructure data in real-time, spotting the subtle signals that often precede system failures. Empirik’s software hooks into existing monitoring tools, scanning logs, metrics, and traces to detect anomalies that human teams might miss. Sequoia’s incubation team, which focuses on high-potential early-stage companies, led the round, with additional backing from angel investors.

The push for proactive IT reliability comes as businesses increasingly rely on complex, distributed systems. Downtime isn’t just costly—it erodes trust, disrupts revenue, and damages reputation. Traditional monitoring tools alert teams *after* something breaks, leaving engineers scrambling to figure out what went wrong. Empirik aims to flip that model by predicting failures before they escalate. Its approach mirrors the shift in software engineering, where tools like Cursor use AI to assist developers in real-time rather than waiting for bugs to surface.

The startup’s technology builds on advances in observability and AI-driven incident management. Empirik’s platform doesn’t just flag issues—it ranks them by severity and suggests fixes, reducing the cognitive load on overstretched IT teams. The $21 million seed funding will go toward product development, hiring engineers, and expanding into cloud-native environments. Early customers include tech-forward enterprises in finance and e-commerce, where downtime is measured in lost transactions and customer churn.

What happens next could reshape how companies handle IT reliability. If Empirik’s predictions hold up, it may force incumbents like Datadog and New Relic to integrate predictive features—or risk losing ground. The broader trend points toward AI that doesn’t just observe systems but actively prevents failures. For industries where uptime is non-negotiable, that shift can’t come soon enough.