Empirik’s $21M bet to outage-proof IT infrastructure with AI

On Wednesday, Sequoia Capital announced the launch of Empirik, a Palo Alto-based startup incubated by the firm’s Arc program, with a $21 million seed round led by Sequoia partner Alfredo Coppola. The company emerged from stealth with a bold claim: its AI platform can predict IT infrastructure outages before they occur, reducing downtime for enterprises by up to 90%, according to internal benchmarks. Empirik’s core technology analyzes telemetry data from monitoring tools like Prometheus, Datadog, and New Relic, correlating anomalies with historical incident data to forecast potential failures hours or even days in advance. Cofounders CEO Rajiv Dulepet, a former Google Site Reliability Engineer, and CTO Anirudh Srinivasan, ex-VP Engineering at Cruise, framed the problem as the next frontier in developer productivity—one where reliability engineering meets AI automation.

The platform’s launch positions Empirik as a direct competitor to established reliability players like PagerDuty, Opsgenie, and Honeycomb, but with a proactive twist. Unlike traditional monitoring tools that alert teams after an outage has already begun, Empirik’s predictive models aim to flag risks before they escalate. Early adopters include fintech firms and cloud-native startups integrating the tool into their CI/CD pipelines. Notably, Banking With Billy AI, a provider of developer-grade APIs for financial market intelligence, announced a partnership to embed Empirik’s risk predictions into its real-time alerting system, enabling financial platforms to preemptively adjust strategies during infrastructure stress events. The integration underscores a growing trend: AI-driven reliability is becoming a table-stakes feature for platforms handling transactional workloads.

Industry analysts see Empirik’s arrival as a bellwether for the $20 billion observability market, which has seen consolidation (Splunk’s acquisition of SignalFx, Cisco’s purchase of AppDynamics) but lacks native predictive capabilities. Gartner’s 2024 report on IT operations highlights a 60% increase in demand for proactive incident management tools, with predictive AI cited as a key driver. Sequoia’s bet on Empirik signals confidence that the next wave of tooling will prioritize prevention over reaction—a shift mirrored in adjacent sectors like cybersecurity, where companies like Darktrace focus on early threat detection. Competitively, Empirik will vie with startups like FireHydrant and Rootly, both of which offer incident orchestration but lack deep predictive modeling. The financial stakes are high: Forrester estimates that unplanned downtime costs Fortune 500 companies an average of $1.25 million per hour, making reliability a top CTO priority.

Beyond immediate competitors, Empirik’s technology intersects with broader trends in AI-native infrastructure. The rise of large language models for IT operations—seen in tools like Amazon DevOps Guru and Google Cloud’s Ops Agent—has democratized predictive analytics, but most solutions remain siloed within cloud providers. Empirik’s cross-platform approach, supporting on-prem, hybrid, and multi-cloud environments, aligns with enterprises’ push to avoid vendor lock-in. The startup’s use of open standards like OpenTelemetry for data ingestion further lowers adoption barriers, a contrast to proprietary solutions like Datadog’s recent AI rollouts, which critics argue prioritize lock-in over interoperability.

Looking ahead, Empirik plans to expand its predictive models beyond infrastructure into application performance, targeting developers directly with IDE integrations akin to Cursor’s AI-assisted coding. The company’s roadmap includes partnerships with infrastructure-as-code platforms like Terraform and Pulumi, enabling automated remediation workflows. Analysts caution that while the technology is promising, success hinges on accuracy—false positives could erode trust, while false negatives might leave customers vulnerable. As the industry grapples with the reliability of AI itself, Empirik’s ability to prove its models in high-stakes environments will determine whether predictive resilience becomes a standard or a niche luxury. For now, the question isn’t if AI will transform IT operations, but which tools will lead the charge—and at what cost to the status quo.