GPU Shortage Returns as Microsoft, Amazon Tighten Supply; AI Startups Face 32% Price Hike and Year-End Queues

Gate News message, April 25 — A GPU shortage is resurfacing as major cloud providers including Microsoft and Amazon concentrate computing capacity toward internal teams and major customers like OpenAI and Anthropic, leaving smaller AI startups facing price increases, extended wait times, and stricter contract terms. Microsoft Azure's sales management has informed staff that GPU wait times for cloud customers are expected to persist through the end of 2026.

Image generation startup Krea, which raised $83 million from investors including Andreessen Horowitz and Bain Capital Ventures, previously rented hundreds of Blackwell chips at $2.80 per hour under a six-month contract. Upon renewal, multiple cloud providers stopped responding to inquiries; Krea ultimately secured chips at $3.70 per hour, a 32% increase, with the contract extended to one year. CEO Victor Perez noted that some vendors simply did not respond, while others insisted on multi-year commitments before negotiating. GPU cloud provider Lightning AI's CEO Will Falcon revealed the company operates 40,000 GPUs online, but approximately 40 customers in queue require a combined 400,000 GPUs, with rental prices rising over 25% within six months.

Microsoft has implemented tiered GPU access management: roughly 1,000 largest customers (Tier 1) receive priority allocation, while smaller clients seeking Blackwell chips must commit to at least 1,000 units for a minimum of one year, with contracts starting at tens of millions of dollars. Pay-as-you-go customers risk losing GPU access if devices remain idle for several hours. Startups participating in Microsoft for Startups' free-tier program have also been warned that insufficient GPU utilization may result in revoked access.

Venture capital firm General Catalyst is surveying its portfolio companies on compute bottlenecks and exploring shared GPU pools or collective negotiation strategies. Some startups are considering direct GPU purchases to bypass queues: oil-industry AI startup Collide plans to spend approximately $500,000 on Nvidia GPUs, leasing data center space to operate independently and avoid uncertainty.