AI Coding Costs Hit $213/Week Per Engineer: Larridin Exposes Hidden Spend

The Gist

• Median engineer spends $213/week on AI coding tokens, with a 10x spread to the 90th percentile
• Annualized, top engineers consume $47K/year in tokens, creating a seven-figure budget gap for 100-person teams
• Larridin connects AI spend to engineering output, revealing where automation ROI plateaus

• AI coding costs average $213 per week per engineer, with the 90th percentile nearing $47,000 annually per engineer in tokens alone.
• AI spend only converts to more output where fluency already exists, meaning there's no universal ideal budget for AI tools.
• Larridin connects AI usage and spend to productivity metrics, helping companies track ROI and optimize AI investments.
• Agent spend is the fastest-growing and least-understood line in most AI budgets, often not mapped to a specific seat or person.
• Companies often make AI procurement mistakes by signing big enterprise licenses without understanding current tool usage, leading to wasted spend.
• Larridin's Router feature optimizes coding tasks by routing requests to lower-cost models when possible, tying savings to output quality.

• Track AI spend per engineer and identify where fluency exists to optimize budget allocation.
• Audit current AI tool adoption before making new enterprise purchases to avoid redundant spend.
• Implement a system like Larridin Router to optimize coding tasks and reduce unnecessary model costs.
• Set up regular reviews of AI spend ROI to identify where budget increases stop yielding additional output.

• $213/week (Average AI coding cost per engineer)
• $47,000/year (AI coding cost for 90th percentile engineers)
• $170/week (Starting AI spend for engineers in the same company)
• $50K/year (Estimated starting enterprise pricing for Larridin)
• $17M (Seed funding raised by Larridin)