# Larridin raises $30M to tell companies whether their AI bill is buying anything

> Its pitch: one engineer burning 65% of an AI budget, and two coding tools 66 times apart in cost per shipped change.

- URL: https://devisdead.ai/news/larridin-raises-30m-to-tell-companies-whether-their-ai-bill-is-buying-anything
- Published: 2026-11-03T10:21:14Z
- Author: DevIsDead Newsroom
- Section: Startups & Money
- Tags: larridin, funding, ai-spend, code-review
- Originally reported by SiliconANGLE (https://siliconangle.com/2026/11/02/vcs-bet-30m-on-larridins-ability-to-help-enterprises-demonstrate-ais-value-for-money/).

Companies have spent two years buying AI seats. Now the finance department wants receipts. Larridin, a startup that measures whether AI spending pays off, has raised a $30 million Series A, as [SiliconANGLE](https://siliconangle.com/2026/11/02/vcs-bet-30m-on-larridins-ability-to-help-enterprises-demonstrate-ais-value-for-money/)'s Mike Wheatley reported.

Pelion Ventures led the round, with existing backers Andreessen Horowitz and Gradient Ventures joining new investors such as Cougar Capital, Alumni Ventures, W23, GFT, Thayer and DG Daiwa Ventures. Larridin has now raised $47 million in total.

## The problem it sells against

Chief executive Russ Fradin's pitch is that the usual adoption numbers, seat licenses and login counts, say almost nothing about impact. One example comes from a customer: its own survey said 92% of its engineers used AI regularly, which looks great on a slide. Larridin's data showed a single engineer was responsible for 65% of all its AI spending.

The other example is cost per result. Larridin says it found a 66-fold difference in cost per shipped code change between two AI coding tools that different teams were using for similar work. The customer cancelled the expensive one and, by Larridin's account, lost nothing in output.

## The review bottleneck, again

Larridin also published its Engineering AI Impact Report, which surveyed over 1,000 developers. The share of pull requests written with AI help climbed to 62% in September, up from 35% in April, while median code review wait times grew by more than 50% over the same months.

If that sounds familiar, it is. LinearB's numbers [said much the same thing](/news/ai-doubled-merged-code-review-is-where-it-piles-up-linearb-s-data-says) this week: AI makes more code, and it piles up in front of humans who still have to review it. Two vendors with different data are telling the same story, which makes it harder to dismiss.

## Benchmarks vs. vibes

Some skepticism is in order. The report comes from a company whose product is the cure for the problem the report describes, and the anecdotes are its own customers' stories, chosen by Larridin. Neither the survey method nor the definition of "value" is spelled out in the source.

Still, the questions are the right ones. Platform teams are already being asked what the token bill buys. Measuring cost per merged change, review time, and who actually uses the tools is more honest than counting logins, whoever ends up doing the measuring.

Larridin lists SurveyMonkey, Vertiv and Gainsight among its customers and plans to spend the money on sales and engineering.

**[Scoreboard](/scoreboard) update:** Humans +1. Finance teams start asking AI for receipts, and one tool turned out to cost 66 times more per change.

Stay undead.