ME News reports that on April 25 (UTC+8), according to monitoring by Beating, multiple AI startups have reported that cloud providers such as Microsoft and Amazon are redirecting GPU computing power toward their internal teams and major clients (OpenAI, Anthropic), leaving smaller customers facing price hikes, extended wait times, and stricter contract terms. Microsoft Azure’s sales leadership recently informed employees that GPU wait times for cloud customers are expected to persist until the end of 2026. Specific example: Krea, an image-generation startup that raised $83 million with investors including Andreessen and Bain Capital Ventures, rented hundreds of Blackwell chips at $2.80 per hour six months ago under a six-month contract. Upon renewal, multiple cloud providers went unresponsive; eventually, Krea secured the same hardware at $3.70 per hour—a 32% increase—with the contract extended to one year. CEO Victor Perez noted that some providers simply did not respond, while others only agreed to negotiate under three-year contracts. Will Falcon, CEO of GPU cloud provider Lightning AI, revealed that while his company currently has 40,000 GPUs online, demand from approximately 40 waiting clients totals 400,000 GPUs, with rental prices rising over 25% in just six months. Microsoft has implemented a tiered access system for GPUs: approximately 1,000 top-tier clients (Tier 1) receive priority allocation. Smaller customers seeking Blackwell chips must commit to renting 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 having their GPU access revoked if their instances remain idle for several hours. Startups participating in Microsoft’s “Microsoft for Startups” free credit program have also been warned that failure to fully utilize allocated resources may result in immediate revocation of GPU access. General Catalyst partner Hemant Taneja has distributed surveys to portfolio companies to assess compute bottlenecks and is exploring options such as shared compute pools or collective negotiation on behalf of portfolio firms. Some startups are now considering purchasing GPUs directly: AI startup Collide, focused on the oil industry, plans to spend approximately $500,000 to buy NVIDIA GPUs outright and rent data center space to operate them independently, thereby avoiding waitlists and uncertainty. (Source: BlockBeats)
GPU Shortage Resurfaces: Microsoft and Cloud Providers Restrict Supply, AI Startups Face 32% Price Increase and Year-Long Delays
KuCoinFlashShare
Summary
GPU demand continues to outpace supply, pushing support and resistance levels for AI startups to new extremes. According to MetaEra, Microsoft and cloud providers are reserving GPUs for internal teams and top clients, forcing smaller firms to face 32% price increases and wait times extending into 2026. Krea, backed by Andreessen and Bain Capital, experienced cost surges after contract renewal. Lightning AI’s CEO states that demand exceeds supply by a factor of ten. With the Ichimoku cloud signaling persistent bottlenecks, Microsoft now requires small clients to commit to at least 1,000 GPUs for a full year. Some startups are shifting to direct GPU purchases to circumvent these constraints.
Source:Show original
Disclaimer: The information on this page may have been obtained from third parties and does not necessarily reflect the views or opinions of KuCoin. This content is provided for general informational purposes only, without any representation or warranty of any kind, nor shall it be construed as financial or investment advice. KuCoin shall not be liable for any errors or omissions, or for any outcomes resulting from the use of this information.
Investments in digital assets can be risky. Please carefully evaluate the risks of a product and your risk tolerance based on your own financial circumstances. For more information, please refer to our Terms of Use and Risk Disclosure.