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

NVIDIA's launch of a new cloud service and revenue-sharing program represents a strategic move to lower barriers for AI startups seeking computational resources. This initiative demonstrates NVIDIA's ... Read original →

NVIDIA’s cloud-based compute access and revenue-sharing model is less about empowering startups and more about locking in the foundational layer of the AI stack. Technically, soaring costs of 3nm EUV fabrication are decoupling chip design from deployment—NVIDIA leverages software-defined hardware to transform GPUs from products into services, forcing EDA, advanced packaging, and HBM supply chains to align with its architecture. On compliance, tightening U.S. export controls compel global AI firms to reconfigure datacenter footprints; this model mitigates some entity-list exposure but adds geopolitical operational overhead. As AMD and Intel accelerate AI accelerator rollouts and foundry capacity in Taiwan, China shifts toward HPC, NVIDIA counters with business-model moats. Within 18 months, expect a surge of fabless, serverless AI startups—and a semiconductor industry pivot from selling chips to sharing revenue.