Autonomous location reliability technology startup Traversal secures $120 million in funding at a $1 billion valuation

Traversal Inc., a startup trying to automate site reliability engineering tasks with artificial intelligence agents, appears to be moving closer to that goal after raising $120 million in new funding, increasing its valuation to $1 billion.

The newly minted unicorn’s round was led by Kleiner Perkins, Redpoint Ventures and Khosla Ventures, three of Silicon Valley’s best-known venture capital firms, whose names should give the company the credibility it needs to accelerate its growth plans. In addition to Battery Ventures and Amex Ventures, Sequoia Capital – another big VC name – also participated in the round.

AI may have made the job of software developers much easier than it used to be, but it has only made life more difficult for the site reliability engineers tasked with maintaining the infrastructure to run that code. The influx of AI-generated code is putting infrastructure engineers under enormous pressure and finding it difficult to keep up.

According to a report from Splunk Inc. and Oxford Economics earlier this year, the cost of business downtime has risen more than 50% in the last two years, depriving businesses worldwide of up to $600 billion annually. That’s because software developers have increased their code output by about 180% over the same period, meaning those poor infrastructure engineers are being forced to work harder than ever to keep up with the growing output.

This is where Traversal comes to the rescue, which comes to the rescue with a promising autonomous site reliability engineer designed to help human specialists diagnose, troubleshoot and prevent problems in the information technology infrastructure. Instead of just standing guard and notifying human engineers when an outage occurs, Traversal actively investigates any problems that arise and does its best to fix them to ease their tremendous workload.

Co-founder and CEO Anish Agarwal said there was an urgent need for a specialized AI location reliability engineer as not even the best general-purpose large language models could prove they were up to the task. He points to the ORCA benchmark that his company developed together with researchers at Cornell Tech and Columbia University. It turns out that even the most advanced boundary models barely achieve 25% accuracy on SRE tasks of moderate difficulty, and their hallucination rates are often above 40%.

“Site reliability engineering is one of the most difficult problems in applied AI, but with every company’s software environment being different and downtime costing companies up to $600 billion per year, the stakes are clearly enormous,” Agarwal said.

Traversal’s AI far exceeds what the best boundary models can achieve. It is designed for use in enterprise IT environments, where it analyzes massive amounts of operational telemetry, including logs, metrics and traces, to provide a baseline for normal behavior. When problems arise, it uses causal machine learning to determine the root cause of the problem and then works out a solution, so all the human engineer has to do is accept the recommendation to restore normal operations.

Agarwal’s background in causal machine learning is key to the much higher accuracy of Traversal’s AI SRE. It is an AI discipline that combines causal inference with machine learning to identify cause-and-effect relationships rather than relying on statistical correlations. Essentially, Traversal’s models allow you to simulate the impact of various actions, interventions and policy changes on IT infrastructure, allowing you to quickly determine what the likely cause of a problem is.

Agarwal said the company aims to build a “super-intelligent AI SRE” that can finally help companies keep up with the rapid pace of their AI-enabled development teams. And he believes that we are well on the way to achieving this goal. “After two years of developing our technology, we are seeing customers increasingly using Traversal in increasingly complex production environments,” he said.

Over the last year, Traversal has increased its revenue more than nine-fold and the number of monthly active users has increased twenty-fold. Its best-known clients include Pepsi Co., DigitalOcean Holdings Inc., Nubank and Harvey AI Corp.

Traversal’s momentum is impressive, but Agarwal said the company is just getting started. With $120 million in the bank, the company will accelerate hiring across its engineering, go-to-market, and R&D teams as it moves toward its goal of building a super-intelligent AI SRE.

Kleiner Perkins’ Aditya Naganath said Traversal is the first company he’s seen bringing “real intelligence” to complex production environments, and he’s confident there will be huge demand for it. “As AI dramatically accelerates the speed at which software is written and delivered, the complexity of managing production will only increase,” he predicted. “Traversal gives engineering teams the information they need to understand their systems, resolve incidents faster, and work with confidence at scales that simply weren’t possible before.”

Pictures: crossing