The Pulse: a trend of trying to cut back on AI spend within eng departments?

2026-06-11 · Source: The Pragmatic Engineer · Field: Business & Management — Corporate Strategy & Leadership, Operations & Process Management, Artificial Intelligence & Machine Learning · Depth: Intermediate, extended

Summary

Engineering departments are increasingly scrutinizing and cutting back on AI spending, driven by a lack of clear return on investment despite impressive underlying metrics. Uber President Andrew McDonald noted the difficulty in linking AI-driven code commits to tangible productivity gains or useful consumer features, revealing Uber's CTO exceeded the 2026 AI budget by March. This trend is evident across companies: OpenCode sees spiking demand for cheaper inference services, while cutting-edge firms consider "smart" model routing. DoorDash empowers developers with token limits and efficiency training, and a large retirement-savings company imposed GitHub Copilot limits, forcing use of less capable "0x" models like GPT-5 mini, GPT-4.1, and Grok Code Fast 1. A bottom-up focus on AI efficiency is emerging from engineers, suggesting future performance rewards for cost savings.

Key takeaway

For engineering leaders managing AI initiatives, it is crucial to establish clear metrics linking AI tool usage to demonstrable productivity gains and useful feature delivery. Implement robust cost management strategies, such as developer-level token limits and model routing, to control escalating AI budgets. Prioritize internal knowledge sharing on efficiency and consider rewarding engineers for significant cost savings, transforming AI spend from an unquantified expense into a measurable investment.

Key insights

Companies struggle to link high AI spending to direct productivity gains or user-facing feature delivery, prompting cost optimization efforts.

Principles

• AI's direct ROI is currently elusive.
• Unjustified AI spend is unsustainable.
• UI transitions to AI risk conversion rates.

Method

Implement developer-level token limits, encourage "smart" model routing based on use case, and foster internal knowledge-sharing for AI efficiency.

In practice

• Set monthly AI token usage limits.
• Route AI tasks to cheaper models.
• Reward engineers for AI cost savings.

Topics

• AI Cost Management
• Engineering Productivity
• Large Language Models
• AI Budgeting
• Model Routing
• Developer Efficiency

Best for: CTO, AI Architect, MLOps Engineer, Director of AI/ML, VP of Engineering/Data, Executive

Related on AIssential

• The Pulse: ‘Tokenmaxxing’ as a weird new trend — Developers are inflating AI token usage to meet internal metrics, while companies face challenges managing AI budgets.
• Companies evaluate aggressive AI spending as costs pile up — Aggressive AI spending is straining corporate margins without clear ROI, prompting re-evaluation despite continued investment.
• AI Costs Are Outpacing Marketing Budgets, So How Do You Strategize? — Escalating AI costs, fueled by agentic AI's high token consumption, demand strategic management focused on value and task-appropriate model selection.
• Tech Leaders, Brace Yourselves: AI Costs Will Only Go Up — LLM pricing is rising, necessitating new funding models to scale AI initiatives effectively.
• Does every feature we build with AI need a token budget? — AI features require comprehensive token budgeting across build, run, and maintenance, guided by hypotheses.
• Intelligence Per Dollar — AI model evaluation is shifting from raw performance to cost-efficiency, prioritizing "intelligence per dollar."

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Pragmatic Engineer.