By Hamza Faisal

Traversal, a New York startup building artificial intelligence agents for site reliability engineering, has raised $120 million in a Series B round at a $1 billion valuation. The round was co-led by Kleiner Perkins, Redpoint Ventures and Khosla Ventures, with participation from Sequoia Capital, Battery Ventures and Amex Ventures, the company announced on November 3.

The funding takes Traversal's total raised to $173 million. The company emerged from stealth in June 2025 with $48 million in combined seed and Series A funding, followed by a strategic investment from Amex Ventures. The new round, two and a half times the size of everything it raised at launch, values a company founded in 2023 at ten figures.

From finding outages to fixing them

Site reliability engineering is the discipline of keeping software running: when a service goes down, engineers sift through logs, metrics and traces to find what broke and how to repair it. Traversal's pitch is to hand much of that investigation to AI agents, combining agentic systems with causal machine learning so the software can distinguish a genuine cause from a mere correlation.

At the centre of the platform are two ideas introduced in March 2026: a Production World Model, which turns raw telemetry, code and deployment history into a continuously updated picture of a production system, and a Causal Search Engine built on top of it. Incident Workers, generally available since September 2026, investigate the moment an anomaly triggers, while Alert Workers, in public beta since the same month, triage incoming alerts.

The company says the approach matters because general-purpose AI agents are weak at this job. On ORCA-bench, a benchmark it built with Cornell Tech and Columbia University, the best of five frontier agents managed only 25.3 percent accuracy on medium-difficulty incidents, falling to 10 percent when reports were vague.

Growth claims investors are buying

Traversal reports that its annual recurring revenue has grown ninefold year to date, with usage up nearly 100-fold. Customers ran more than 150,000 investigations last quarter, it says, with the platform supporting more than 4,000 engineers across more than 1,000 major incidents and 50,000 minor incidents.

It also cites greater than 90 percent root-cause accuracy on high-impact incidents and an average 85 percent reduction in mean time to recovery on critical incidents. Those figures come from the company itself and have not been independently verified. Still, the customer list it names, including DigitalOcean and PepsiCo, suggests the product is being tested where downtime is expensive: Traversal cites research putting the cost of unplanned outages at $600 billion for Global 2000 companies in 2026, up 50 percent in two years.

The outer loop of the AI software cycle

The round is a sign of where investor appetite is moving. Much of the AI funding boom has gone to tools that write code faster; Traversal is betting on the other half of the cycle, what it calls the outer loop: keeping that code running once it ships. As AI-written software ships more quickly, the volume of incidents and alerts grows with it, and engineers on call cannot investigate each one by hand.

Traversal describes its ambition as self-driving production: systems that detect problems, understand their causes, take corrective action, verify the fix, and prevent recurrence. Today's product still works alongside human engineers. But the $1 billion valuation shows investors will pay now for a company building the operations layer for AI-written software, not another coding assistant. Whether its agents can resolve incidents reliably in complex enterprise environments, not just diagnose them, is the test that will decide if the valuation holds.