Traversal Inc. has raised $120 million in a funding round that values the site reliability engineering startup at $1 billion. The company says it will use the capital to increase hiring across engineering, go-to-market and research and development as it continues building an AI system intended to investigate and help resolve production infrastructure problems.
The round was led by Kleiner Perkins, Redpoint Ventures and Khosla Ventures. Sequoia Capital, Battery Ventures and Amex Ventures also participated. Traversal’s platform is designed to support human site reliability engineers (SREs), rather than simply monitor systems and send alerts when something goes wrong.
From monitoring to incident investigation
Traversal’s software is deployed within enterprise IT environments and analyzes operational telemetry, including logs, metrics and traces. The company says the system uses those signals to establish a baseline for normal behavior and, when an incident occurs, applies causal machine learning to identify potential root causes and recommend a response.
That approach is intended to go beyond conventional alerting. Instead of handing an incident to an engineer with only a notification, Traversal says its AI investigates the problem and develops a proposed fix. A human engineer can then accept the recommendation to restore service. The source material does not detail the range of actions the system can execute autonomously, so the described workflow centers on diagnosis and recommendations subject to human acceptance.
Co-founder and Chief Executive Anish Agarwal argues that SRE work presents challenges that general-purpose language models are not yet equipped to handle reliably. He cited ORCA-bench, a benchmark Traversal developed with researchers from Cornell Tech and Columbia University. According to the company’s account of the benchmark, leading frontier models achieve less than 25% accuracy on medium-difficulty SRE tasks, while hallucination rates can exceed 40%.
Causal machine learning is central to Traversal’s stated technical approach. Unlike methods that rely primarily on statistical correlations, causal inference seeks to identify cause-and-effect relationships. Agarwal says this enables the system to consider how interventions or policy changes could affect infrastructure, helping it assess possible causes and responses. The company has not provided details in the supplied material about the benchmark’s full methodology or independent validation of its platform’s performance.
Pressure on reliability teams
The company is positioning its product against the operational demands created by faster software development. More code being produced can mean more changes for infrastructure and reliability teams to monitor, troubleshoot and maintain. Traversal’s pitch is that an SRE-focused AI system can help those teams keep pace by taking on portions of incident investigation.
Agarwal pointed to a report by Splunk and Oxford Economics that found enterprise downtime costs had risen by more than 50% over two years and could reach $600 billion annually worldwide. The report also estimated that software developers increased code output by around 180% over the same period. Those figures provide context for Traversal’s market proposition; they are findings attributed to the report, not measurements of Traversal customers or its product.
Traversal says customer use has broadened across increasingly complex production environments. Over the past year, the company reports that revenue grew more than ninefold and monthly active users increased 20 times. It named PepsiCo, DigitalOcean, Nubank and Harvey AI among its customers. The company did not disclose absolute revenue, user totals or customer-specific results in the supplied material.
Funding to support expansion
The new financing gives Traversal resources to scale its teams while it pursues what Agarwal has described as a “superintelligent AI SRE.” The phrase reflects the company’s stated ambition, rather than a technical capability independently established by the funding announcement. The company’s near-term plans, as described, focus on hiring across product engineering, commercial operations and research and development.
Kleiner Perkins partner Aditya Naganath said the company stood out to him for bringing intelligence into complex production environments. He also argued that AI-assisted software development will increase the complexity of managing production systems, strengthening demand for tools that help teams understand and resolve incidents.
Traversal’s approach places it in the expanding category of AI-assisted operations software, where the value proposition depends on both useful diagnosis and dependable recommendations in environments that differ from one enterprise to another. Its funding and reported customer growth signal investor and market interest, while the practical test will be how reliably the platform supports engineers across real production incidents.