Empirik just launched out of Sequoia Capital with $21 million in seed funding, and its pitch is simple: catch tech outages before they knock systems offline. According to TechCrunch AI, the startup spun out as an independent company on Tuesday, backed by Sequoia, Canapi, and Alumni Ventures. It’s the latest bet that AI can do for infrastructure engineers what coding assistants already did for developers.
Who’s behind it
The founders aren’t newcomers. Avon Puri ran infrastructure at Rubrik and VMware for over a decade before becoming Sequoia’s chief digital and information officer in 2020. He teamed up with fellow Sequoia IT leader Sudheer Dhurjati about three years ago, right as large language models started showing real muscle.
Their insight, as detailed in TechCrunch AI: stop reacting to failures after they happen and start predicting them. Sequoia liked the idea enough to incubate the company in 2023, then recruited Kartik Chandrayana, former Quantum Metric CPO and Salesforce observability VP, as CEO earlier this year.
What Empirik actually does
Most observability tools tell you something broke after it breaks. Empirik works the other way. It tracks changes across your systems and maps the ripple effects a single change could trigger across the whole environment.
Think of it as an autonomous traffic cop for infrastructure changes:
- Waves through low-risk changes automatically
- Sets guardrails on bigger ones
- Flags the most dangerous updates for a human to review
The goal is to hand off routine troubleshooting so DevOps and site reliability engineering teams can spend their time on work that matters more. “There has always been a lot of money spent in keeping systems up and running,” Sequoia partner Bogomil Balkansky told TechCrunch, arguing that most existing tools don’t grasp how complex system dependencies actually connect.
Why this matters
What stands out here is the framing. Chandrayana wants Empirik to be for infrastructure engineering what agentic AI became for software development. “What agentic AI did for software, Empirik wants to do for infrastructure engineering,” he told TechCrunch. The comparison to Cursor and Claude Code is deliberate, and it’s a smart read on where the market is heading.
Here’s the pressure point. AI is speeding up how fast code gets written and shipped. Every one of those changes lands on infrastructure that humans still have to babysit. The faster developers move, the more system changes pile up, and the harder it gets for SRE teams to keep track of what might break. A tool that predicts the blast radius of a change before it goes live fits neatly into that gap.
The old status quo was reactive. Alerts fire, dashboards light up, engineers scramble to trace the root cause while the clock runs. Empirik is trying to flip that into something proactive, closer to prevention than firefighting.
Early traction
This isn’t just a slide deck. Since a quiet launch earlier this year, Empirik says it has signed customers ranging from startups to Fortune 500 names, including S&P Global, Guardant Health, and a major consumer packaged goods company. Landing enterprise logos that fast is a real signal that the problem is painful enough for big companies to pay for a fix.
On competition, Balkansky claims Empirik sits in a category of its own for now. He positions it as a complementary layer that sits alongside AI SRE platforms like Resolve and Sequoia-backed Traversal, rather than going head to head with them.
What to watch next
A few things worth tracking:
- Whether “predict before it breaks” holds up at scale across messy, real-world environments
- How much autonomy those enterprise customers actually give the tool versus keeping humans in the loop
- Whether rivals in the observability and AI SRE space start building the same change-tracking layer
The seed round buys Empirik room to prove the model works beyond its early customers. If it does, expect the observability crowd to move fast in the same direction. For now, the company has money, marquee logos, and a clear thesis about where infrastructure work is going. You can find the full details at the original source.