Nvidia Targets $1 Trillion AI Chip Market

This bold outlook, unveiled at the GTC developer conference, aims to allay investor concerns regarding Nvidia's growth trajectory and its strategy of reinvesting profits into the AI ecosystem. The company is aggressively expanding its focus beyond AI model training, where its GPUs have dominated, into the rapidly growing "inference computing" market—the process of real-time query answering. To solidify its position in inference, Nvidia introduced a new central processor (Vera CPU) and an AI system leveraging technology licensed from Groq for $17 billion.

This strategic move addresses increasing competition from CPUs by Intel and custom processors from Google in the inference space. Huang detailed a two-step inference process, with Nvidia's Vera Rubin chips handling "prefill" and Groq's chips managing the "decode" stage.

The company also outlined its future Feynman roadmap and introduced NemoClaw for autonomous AI agents, demonstrating a comprehensive approach to AI infrastructure. Analysts view these announcements as reinforcing Nvidia's leadership and signaling the AI industry's transition from experimentation to large-scale deployment, despite the stock's modest post-announcement gain.