WTF is Anthropic doing???

2026-04-23 · Source: Matthew Berman · Field: Business & Management — Corporate Strategy & Leadership, Artificial Intelligence & Machine Learning, Project & Product Management · Depth: Intermediate, extended

Summary

Anthropic is facing significant user frustration and operational challenges stemming from a strategic miscalculation regarding compute investment and inconsistent communication. The company recently removed Claude Code from lower subscription tiers, then partially rolled it back, causing confusion. This follows a core decision by CEO Dario Amodei to not aggressively invest in compute capacity, fearing bankruptcy if demand growth projections were not met. However, demand for AI, particularly for agentic use cases, has surged, leaving Anthropic with insufficient compute to meet user needs. This has led to a series of "carrot and stick" policies, including off-peak usage incentives and stricter session limits for power users, alongside opaque communication regarding policy changes and token usage, further alienating its user base. Competitor OpenAI is actively capitalizing on Anthropic's missteps, offering more generous usage limits and acquiring key projects like OpenClaw, while maintaining higher uptime and clearer communication.

Key takeaway

For CTOs and VPs of Engineering evaluating AI platform partnerships, Anthropic's current struggles highlight the critical importance of a vendor's compute strategy and transparent communication. Your teams risk operational disruptions and policy uncertainty if a provider cannot scale or clearly articulate usage terms. Prioritize partners like OpenAI or Google who demonstrate robust compute infrastructure and consistent, clear policies to ensure long-term stability and avoid unexpected costs or service limitations.

Key insights

Underestimating AI demand and compute needs can severely impact user trust and market position.

Principles

• Clear communication builds trust.
• Compute capacity is critical for AI growth.

Method

Anthropic's "flywheel" model involves building coding models, selling to enterprise, acquiring coding data, and using that data to train next-gen models, fueled by compute investment.

In practice

• Prioritize compute infrastructure for scaling AI services.
• Maintain transparent communication on product changes.
• Monitor competitor actions for market positioning.

Topics

• Anthropic Policy Changes
• Compute Shortage
• AI Demand Forecasting
• OpenAI Competitive Strategy
• Claude Code

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

Related on AIssential

• The Pulse: Did capacity shortages turn Anthropic hostile to devs? — Anthropic's rapid growth led to capacity issues, prompting a strategic compute partnership with rival xAI.
• 😸 The feud running the AI industry — The core ideological split between OpenAI and Anthropic drives major AI industry decisions.
• Anthropic takes the path of most resistance — Effective AI deployment requires robust operational infrastructure and strategic communication, not just powerful models.
• Wtf Anthropic — Anthropic's compute miscalculation and opaque policies are alienating users, while OpenAI aggressively captures market share.
• 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.
• [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.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Matthew Berman.