Nvidia Is Making it Easier for AI Startups to Get Compute Power With a New Cloud and Revenue-Sharing Program - Yahoo Finance

NVIDIA has launched a new initiative aimed at helping artificial intelligence (AI) startups access high-performance computing resources more easily. Through a revenue-sharing and credit-support model,... Read original →

NVIDIA’s new cloud and revenue-sharing program is less about democratizing AI than entrenching its GPU stack as the de facto infrastructure layer. Technically, tight integration between Grace Blackwell GB300 and DSX accelerates demand for 3nm/EUV nodes, intensifying foundry pressure on TSMC (Taiwan, China) while marginalizing general-purpose CPUs in data centers. Regulatory scrutiny looms: U.S. and EU antitrust authorities may view its $30B OpenAI stake as circular financing that distorts competition. Rivals will respond—AMD could leverage MI300X partnerships with hyperscalers, while Intel might undercut with Gaudi 4 plus foundry bundling. Within 18 months, most AI startups will find themselves architecturally locked into NVIDIA’s ecosystem, creating a ‘compute colonialism’ effect: lower entry barriers today, but irreversible dependency tomorrow—solidifying NVIDIA’s role as quasi-utility in the AI semiconductor order.