From Training AI to Designing Chips: Nvidia's Next Platform Play

Core Insights - Nvidia's CEO Jensen Huang highlighted the upcoming delivery of the Rubin platform at CES 2026, which is generating significant excitement among analysts due to its innovations in physical AI [1][2] Performance and Efficiency - The Rubin platform consists of six separate Nvidia chips and is expected to lower the cost per token for AI inference workloads by up to 90% compared to the Blackwell platform, enhancing performance and efficiency [2] - Data center operators are expected to benefit greatly from the Rubin platform, which will have substantial impacts on performance and efficiency at scale [3] Market Position and Growth - Nvidia currently holds a market capitalization of approximately $4.5 trillion and anticipates major companies like Microsoft, Oracle, Amazon, Alphabet, and CoreWeave to be among the first to adopt the Rubin systems [3] - The transition from a chip supplier to an end-to-end platform provider is seen as a crucial long-term growth engine for Nvidia [4] Revenue Projections - Nvidia projects that combined sales from the Rubin and Blackwell platforms could reach around $500 billion in 2026, representing a significant increase from its recent quarterly revenue of $57 billion and trailing-12-month revenue of $187 billion [5] - The company expects to generate $65 billion in revenue for the fourth quarter of fiscal 2026, ending January 25 [7]