Undo secures €31 million to convey runtime context to AI-assisted software program engineering | EU-Startups


Cambridge-based Undo, a scale-up centered AI-powered root trigger evaluation, right now introduced the shut of a €31 million ($37 million) funding spherical to speed up growth efforts and considerably increase its world market attain.

The spherical was led by Elsewhere Companions.

We’re  forward of the curve. Undo has spent years constructing deterministic, program recording know-how for code failure runtime visibility, which has turn out to be completely important with the rise of AI,” says Undo Founder and CEO Greg Legislation.

This funding permits us to speed up at precisely the fitting second – embedding Undo into AI workflows, scaling our industrial attain, and making certain we’re an important a part of how engineering groups function on this new, AI-first world. We sit up for working alongside the Elsewhere crew to ease the following period of software program engineering issues for corporations around the globe.

Comparable 2026 exercise factors to sustained funding exercise across the operational layers of AI-assisted software program growth: software program provide chain management, AI-agent safety and governance, CI/CD, cloud operations, manufacturing infrastructure, and developer tooling.

Undo’s spherical sits closest to this group of corporations as a result of it addresses an analogous enterprise concern: how engineering groups can preserve reliability, visibility and management as AI-generated or AI-assisted code will increase system complexity.

The UK comparability is especially related, with Cloudsmith, Geordie AI, Trent AI, CodeWords, Overmind and Toyo all receiving funding in 2026, indicating home exercise round safe and dependable AI-enabled software program infrastructure.

AI is making code unmanageable – introducing code that engineers can not perceive, belief, or debug. So whereas AI helps them generate extra code, a few of it’s poorly understood, poorly structured, and of questionable high quality. Programs turn out to be stuffed with unknowns, making them unstable and rising the chance of outages, safety breaches, and buyer escalations,” notes Elsewhere Working Companion Rod Favaron.

Undo ensures engineering groups can successfully function advanced programs in an AI-first world and supply the important runtime context for enterprise-grade, AI-assisted software program.

Based in 2012, Undo permits coding brokers to unravel advanced issues on advanced codebases, enabling absolutely automated root trigger evaluation – throughout growth, check and manufacturing.

Undo says they fill a important hole by giving AI brokers the runtime context wanted to reliably diagnose advanced software program points in multifaceted programs. By capturing full execution historical past (how code truly behaves when working) into self-contained recordings, Undo permits AI brokers to carry out correct root trigger evaluation and ensures even AI-generated code stays comprehensible and maintainable.

Irrespective of how good AI fashions get, the corporate outlines that outcomes are restricted by how good the context is: AI = mannequin + context. Runtime context tells the mannequin what the programme did, moderately than simply what the code says, so the fashions carry out considerably higher.

In keeping with the corporate’s benchmarks on a variety of advanced bugs:

  • The most recent fashions are capable of establish the root-cause of simply 38% with out Undo; whereas with Undo’s runtime context that will increase to 92%.
  • In instances the place fashions can remedy an issue with out Undo, when the identical downside is solved with Undo fewer tokens are used.
  • Clients report completion of root-cause evaluation 100x quicker with Undo than earlier than.

High quality is extraordinarily essential at Palo Alto Networks, and we can not afford to depend on guesswork. The toughest – and costliest – bugs in multi-million-line codebases stay in runtime state and usually are not captured by logs or different options,” says Suresh Sangiah, Senior VP Engineering at Palo Alto Networks. “Undo gives the visibility wanted to catch and proper errors earlier than they turn out to be an operational downside for our prospects, enabling automated root trigger evaluation. Undo typically autonomously finds the foundation causes in minutes.

The corporate plans to considerably scale its product growth, buyer assist, and go-to-market groups throughout the USA and Europe to assist the growth and adoption of its options as software program engineering wants evolve.