Nvidia Offers Revenue Sharing Model for Aspiring AI Startups - Bloomberg.com

NVIDIA's introduction of a revenue-sharing model for emerging AI startups represents a significant strategic shift in the semiconductor industry's approach to ecosystem development. This move reflects... Read original →

NVIDIA’s revenue-sharing model effectively positions its GPU stack as the de facto 'operating system' of AI, locking startups into a vertically integrated ecosystem. This accelerates software-stack consolidation, marginalizing open alternatives like RISC-V in training workloads. For TSMC (Taiwan, China), near-term wafer demand gains come with heightened geopolitical exposure—if U.S. export controls tighten on advanced nodes, NVIDIA-dependent startups may scramble to diversify foundry sources. AMD and Intel are unlikely to mirror pure revenue-sharing; instead, they’ll likely counter with hybrid IP licensing plus foundry subsidies to protect margins. Over the next 18 months, this strategy will spawn a wave of 'NVIDIA-native' AI firms but risks systemic fragility: a breakthrough in photonic or in-memory computing could abruptly invalidate today’s ecosystem lock-in.