Nvidia Says It Will Take a Cut of Some Customers’ Cloud Revenues - The Information

NVIDIA has announced that it will take a cut from some customers' cloud revenues, signaling a strategic expansion into cloud services. This move underscores NVIDIA's growing influence in AI computing ... Read original →

NVIDIA’s shift from chip sales to taking a cut of cloud revenues marks a fundamental reconfiguration of the AI compute value chain. Technically, this forces deeper CUDA stack lock-in, amplifying reliance on 3nm/EUV chips and cementing TSMC (Taiwan, China) as the indispensable foundry partner. Regulatory risks loom large—revenue-sharing across jurisdictions may trigger data localization scrutiny in the U.S. and EU, raising compliance overhead. Competitors will respond aggressively: AMD could accelerate MI300X adoption with ROCm integration, while Intel might bundle chips with IFS foundry services. Within 12–24 months, a 'Hardware-as-a-Service' norm will emerge, pricing out smaller cloud providers from high-end AI training and accelerating market concentration among hyperscalers.