Larridin Blog

AI Coding Spend per Engineer: 2026 Benchmark Data

Written by Larridin | Nov 16, 2026

The Larridin Benchmark puts the median billed AI coding spend at $213 per engineer per week. Our CTO, Ameya Kanitkar, shared that number on LinkedIn with an important catch: more spend doesn't automatically mean more output.

He explained his thinking on LinkedIn:

Read Ameya's post on LinkedIn

Ameya Kanitkar, co-founder and CTO of Larridin, on LinkedIn:

The Larridin Benchmark data indicates that the median billed spend for AI coding is $213 per engineer per week (approximately $920 per month), reaching ~$3640 per month at the 90th percentile.

And more spend does not always correlate with more engineering output but it does once, you bring in AI Fluency in the picture!

Why This Matters if You Own an Engineering Budget

AI coding is a real line item now, and sooner or later finance will ask whether your number is high or low. A benchmark helps you answer that. What it can't do on its own is tell you whether the money is buying anything.

The Numbers

Here's where billed AI coding spend lands in the Larridin Benchmark:

  • Median: $213 per engineer per week, or about $920 per month
  • 90th percentile: about $3,900 per month

The original LinkedIn post above used approximately $3,640 per month. The published benchmark now lists $911 per week at the 90th percentile, or approximately $3,900 per month; the figures here follow that benchmark.

Spend and Output Only Line Up When Fluency Is There

The post Ameya reshared from Jason Saltzman digs into why. The two AI-native groups came from the same company, using the same tools and prices. The difference in output was associated with AI fluency.

A few findings stand out:

  • Spend and output only correlate for AI-native engineers
  • At equal spend, deeply AI-native engineers ship about twice as much as partial adopters
  • Partial adopters see real returns, but their marginal payoff drops by about half once spend passes roughly $600 per week
  • The deeply AI-native group hasn't hit a point yet where more spend becomes less productive
  • For low-AI teams, extra dollars buy activity without output

Today the gap between AI-native engineers and everyone else is about 2x at equal spend, and the data suggests it isn't done widening.

What to Do With This

A single spend target won't get you far. A few more useful moves:

  • Compare spend per engineer against the median and the 90th percentile to spot outliers in both directions
  • Look at spend next to output and AI fluency, so you can tell productive heavy users from expensive low-output ones
  • Before cutting anything, find out what the spend produces; Ameya is wary of hard caps for exactly this reason since they stop useful work as easily as wasteful work
  • Invest in fluency because that's where the return shows up

Closing Thought

Spend is an input. An engineer spending $213 a week could be one of your most productive people or one of your least, and the number alone can't tell you which. If you pair spend with output and fluency, it starts to mean something.

Ameya shares this kind of practical, numbers-first stuff on LinkedIn pretty regularly. If you want more of it in your feed, follow Ameya Kanitkar on LinkedIn. He's co-founder and CTO of Larridin.

On larridin.com: