The Pulse: Did capacity shortages turn Anthropic hostile to devs?

2026-05-07 · Source: The Pragmatic Engineer · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Fundamental Awareness, quick

Summary

Anthropic has recently drawn developer criticism for perceived model degradation and restricted Claude Code access, potentially due to undisclosed capacity constraints despite securing compute from SpaceX. Concurrently, Amazon has reversed its ban on Claude Code and Codex usage, aiming to enhance its Kiro coding agent. Meta is reassigning 20-40% of engineers to data labeling tasks, raising questions about job security and efficiency. A new trend towards small "AI-forward" teams of 5-10 developers is emerging at Meta and Amazon, with CEOs suggesting these outperform larger teams, though implications for "excess" staff and task scope remain unclear. Additionally, GitHub is implementing significant pricing changes for Copilot, shifting Pro and Pro+ plans to roughly a 3x increase on annual renewal and other plans to unpredictable API token-based pricing starting June 1.

Key takeaway

For CTOs and entrepreneurs evaluating AI tool adoption and team structuring, be aware of the shifting landscape. Anthropic's capacity issues highlight the fragility of AI service access, while GitHub's pricing changes underscore cost unpredictability. Consider Amazon's move to integrate external AI tools and the trend towards smaller, focused AI teams when planning your own development strategies and resource allocation.

Key insights

Big Tech is navigating AI capacity, developer relations, team structures, and pricing model shifts.

Principles

• AI capacity dictates developer access
• Smaller teams can drive AI innovation

In practice

• Monitor AI provider capacity signals
• Evaluate small team structures for AI projects

Topics

• Anthropic Developer Relations
• AI Model Capacity
• AI Coding Tools
• GitHub Copilot Pricing
• Data Labeling

Best for: CTO, Entrepreneur, Director of AI/ML, VP of Engineering/Data, AI Product Manager

Related on AIssential

• TAI #197: Anthropic Turned the OpenClaw Demand Signal Into a Product — Agentic AI advancements are rapidly increasing compute demand, straining a constrained global AI supply chain.
• The Debate Over Anthropic’s New Product: Price or Existential Dread? — AI agent pricing and capabilities are sparking an existential re-evaluation of traditional software development workflows and costs.
• Claude just changed overnight — Anthropic restricted third-party access to subsidized tokens, citing cost and optimization, sparking community backlash over perceived ecosystem control.
• Anthropic manager hints that Pro and Max plans are outgrown by today's Claude workloads — AI providers face a compute crunch, leading to service restrictions and plan adjustments for compute-intensive tools.
• [AINews] Anthropic-SpaceXai's 300MW/$5B/yr deal for Colossus I, ARR growth is 8000% annualized — Compute capacity and agent system ergonomics are now as critical as base model intelligence for frontier AI labs.
• The Pulse: AI load breaks GitHub – why not other vendors? — Increased AI-driven load is impacting core developer infrastructure reliability and business models.

Open in AIssential →

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