AI is changing the world. Don't stay behind. Clear summaries, community insight, delivered without the noise. Subscribe to never miss a beat.

The beginning of scarcity in AI

tomtunguz.com

April 16, 2026

1 min read

🔥🔥🔥🔥🔥

57/100

Summary

Technology companies are facing supply chain limits for the first time since the 2000s, leading to a 48% increase in GPU rental prices for Nvidia’s Blackwell chips, now at $4.08 per hour. CoreWeave has raised prices by 20% and extended minimum contracts from one year to three, while OpenAI's CFO, Sarah Friar, indicated that compute shortages are forcing difficult decisions on project priorities.

Key Takeaways

GPU rental prices for Nvidia’s Blackwell chips increased by 48% to $4.08 per hour within two months.
OpenAI CFO Sarah Friar stated that the company is making tough decisions on projects due to insufficient compute resources.
Access to advanced AI models is becoming restricted, with Anthropic limiting its newest model to approximately forty organizations.
The current AI landscape is characterized by rising costs, slower access to models, and a shift towards smaller models and on-premise deployments.

Read original article

Community Sentiment

Mixed

Positives

The current constraints in AI may drive innovation in harness design and small models, leading to more efficient and cost-effective solutions.
The rapid improvement of AI models means that mid-level models can now perform tasks that once required frontier models, potentially lowering costs for many applications.
The ongoing economic pressures could lead to a resurgence in local model deployment, which may democratize AI access and reduce reliance on expensive cloud services.

Concerns

Companies that are heavily reliant on AI may face significant price increases, which could impact their competitiveness and profitability in a saturated market.
The high valuations of AI companies like OpenAI may not be sustainable given their cash burn rates, raising concerns about the long-term viability of their business models.
The assumption of finite hardware production capacity suggests that the current scarcity in AI resources could hinder growth and innovation in the short term.