Nvidia shifts toward AI inference with Vera and Groq systems

Illustration of Jensen Huang presenting Nvidia’s next-generation AI infrastructure beside a rack-scale system at GTC 2026 in San Jose, AI generated. / AI Generated
Nvidia used its GTC 2026 keynote to argue that the next big AI spending wave will come from inference, unveiling new Vera CPU and Groq-based systems while lifting its AI chip revenue opportunity forecast to at least $1 trillion through 2027.











