Empirik, which uses AI to predict and prevent IT outages by analyzing system changes and their effects, raised a $21M seed and spins out from Sequoia Capital
Before joining Sequoia Capital in 2020 as chief digital and information officer, Avon Puri spent over a decade running infrastructure at Rubrik and VMware.
Context & Ripple Effects
Empirik emerges from Sequoia with a founder whose infrastructure experience spans VMware, Rubrik, and Sequoia itself. Its approach applies a predictive-maintenance model already used in industrial settings: Augury used machine signals to anticipate equipment faults, while Empirik focuses on the consequences of changes to enterprise IT systems.
The enterprise AI field has also been moving toward prevention rather than post-incident response. In July, Way Security raised seed funding for AI-driven IAM deployment automation, and Empirical Security raised a Series A for predicting threats from exploited-vulnerability monitoring; Empirik targets the operational-outage side of that broader push.
First-order effects
- Empirik becomes an independent, $21 million-funded company, giving its team resources to develop and sell its change-analysis platform beyond Sequoia.
- Enterprise infrastructure teams gain a vendor focused on forecasting the effects of system changes before those changes trigger outages.
Second-order effects
- Way Security and Empirical Security address adjacent enterprise-risk workflows—identity deployment and vulnerability exposure—while Empirik gives buyers a separate preventive tool for infrastructure change risk.
- For enterprise buyers, outage prevention becomes a product category that can be evaluated alongside security automation rather than only through post-incident monitoring and response.
Third-order effects
- If enterprises adopt tools that model the operational impact of changes before deployment, infrastructure operations may shift from alert-driven remediation toward AI-assisted pre-change decision-making.
- The pattern points to enterprise AI vendors unbundling discrete reliability and risk tasks from broader IT operations platforms, with specialist tools competing for a place in the deployment workflow.
The trend: Enterprise AI is moving from detecting incidents after they occur toward predicting the operational and security consequences of changes before they are deployed.