Pomegra
Traversal raised a $120M Series B at a $1B valuation, co-led by Kleiner Perkins, Redpoint and Khosla, to build AI that diagnoses and fixes production outages.
Key Takeaways
- Traversal closed a $120M Series B at a $1B valuation on November 3, 2026.
- Kleiner Perkins, Redpoint Ventures and Khosla Ventures co-led; total funding is now $173M.
- Traversal reports 9x year-to-date ARR growth and 150,000+ investigations last quarter.
Lead
Traversal, a startup building an AI site reliability engineer (AI SRE), announced a $120M Series B at a $1B valuation on November 3, 2026. Kleiner Perkins, Redpoint Ventures and Khosla Ventures co-led the round. Sequoia Capital, Battery Ventures and Amex Ventures participated, and Kleiner Perkins and Sequoia, both earlier backers, increased their stakes. The company says it will spend the money on research, engineering and go-to-market hiring, and on extending its AI across the software development lifecycle.
What Did Traversal Announce?
Traversal announced a $120M Series B that lifts total funding to $173M. The company emerged from stealth in June 2025 with a $48M Series A led by Sequoia and Kleiner Perkins. It later took a strategic investment from Amex Ventures, the corporate venture arm of American Express. The Series B alone is 2.5 times the size of the Series A.
Traversal did not disclose its prior valuation or its revenue. The only growth figures it published are relative ones. Annual recurring revenue is up 9x year to date, usage is up roughly 100x, and monthly active users are up roughly 20x. Those multiples say little without a base, and an early-stage company can post a 9x gain from a small starting number.
What Does Traversal's AI SRE Actually Do?
Traversal's product investigates production incidents, identifies root causes, proposes or applies fixes, and flags changes likely to cause future failures. It is built around two components the company calls a Production World Model and a Causal Search Engine. The aim is to reason about why a system broke instead of only surfacing correlated alerts.
It competes for the on-call engineer's time. Traditional observability and incident-management tools collect telemetry and route pages to humans. Traversal's pitch is that software should do the diagnosis before a human opens a laptop.
The company reports more than 150,000 investigations in the most recent quarter. It says it supports over 4,000 engineers and has handled more than 1,000 major incidents, more than 50,000 minor ones and more than 5,000 production changes. It also claims customers have realized tens of millions of dollars in return. Named customers include Nubank, Harvey, DigitalOcean, PepsiCo and American Express, along with unnamed Fortune 100 companies. Nubank CEO David Vélez said that five weeks into the rollout, Traversal was already becoming part of how on-call engineers work incidents.
Is the $1B Valuation Justified by the Numbers?
The valuation rests on trajectory, not on disclosed financials. Traversal went from a $48M Series A in June 2025 to a unicorn in about 17 months, a pace now common in AI infrastructure when usage curves are steep. The round also suggests investors treat incident response as a budget line that can shift from headcount and tooling to software.
The caveats are specific. Revenue is undisclosed, and customer ROI figures are the company's own. Production environments are also an unforgiving test bed. An AI that diagnoses incidents with high accuracy is useful. One that applies a wrong fix to a live payment system is a liability, which is why enterprise buyers tend to start with read-only investigation before granting write access. Traversal's count of 5,000+ production changes suggests some customers have moved beyond that stage, though the company did not say how many of those changes were applied autonomously.
The investor mix points to conviction. Three firms co-led, and both prior-round leaders added capital instead of passing. Follow-on participation from existing investors is generally read as a signal that internal metrics match the pitch.
Who Built It?
The founding team comes from research and quantitative finance. CEO Anish Agarwal holds an MIT PhD. Co-founder Raj Agrawal also holds an MIT PhD. Raaz Dwivedi holds a UC Berkeley PhD and is on the faculty at Cornell Tech. Ahmed Lone is a former quant trader at Citadel Securities. Their focus on causal inference, rather than log summarization, is the technical bet that separates Traversal's pitch from a language model wrapped around a monitoring dashboard.
What Comes Next for AI SRE?
The next stage is autonomy. Traversal describes its goal as self-driving production systems, meaning software that detects, diagnoses, fixes and prevents incidents with limited human involvement. Getting there requires customers to extend trust gradually, from investigation to recommended fixes to approved changes.
Expect pressure from two sides. Observability and incident-management vendors are adding their own AI features and hold the telemetry data. Other AI SRE startups are chasing the same enterprise accounts. Traversal's advantage will depend on whether its causal approach produces measurably fewer wrong diagnoses, and whether it can show revenue that supports a $1B price.
Outlook
Traversal enters 2027 with $173M raised, three co-leading investors and an enterprise customer list that includes a Fortune 100 contingent. The growth multiples are strong but lack disclosed bases. The test now is whether customers let the system act on production without a human approving each step, and whether the revenue behind the 9x figure grows large enough to justify the valuation.
