At CES, Nvidia CEO Jensen Huang announced that the next-generation Rubin system is already in full production—six months ahead of the previously guided timeline. The Rubin platform integrates six specialized chips (CPU, GPU, DPU, networking and switching components) into a rack-scale solution designed for AI workloads, promising up to a five-fold improvement in inference performance and 3.5 times the training throughput versus the current Blackwell architecture. With 60% more transistors than Blackwell and a focus on reducing data-center bottlenecks, Rubin is positioned to drive both cost savings and higher utilization in hyperscale AI deployments.

In its most recent quarter, Nvidia reported a 62% year-over-year revenue increase to $57 billion and a 65% rise in net income to $31 billion, maintaining a gross margin north of 70%. The company’s backlog has surpassed $500 billion, covering orders through early 2027. Based on consensus forecasts of approximately $213 billion in revenue for fiscal 2026, a price-to-sales ratio of 28 would imply a $6 trillion market capitalization—roughly a 34% upside from current levels—making Nvidia the favorite to become the world’s first $6 trillion company in 2026, assuming continued AI spending growth and successful Rubin ramp.

Taiwan Semiconductor Manufacturing, Nvidia’s primary chip fab partner, reported Q4 results that exceeded analyst expectations and highlighted sustained strength in AI demand. CEO C.C. Wei stated that conversations with cloud-service providers confirm a robust multi-year AI megatrend, reinforcing confidence that GPU orders will remain elevated. Given TSMC’s visibility into aggregate orders for Nvidia and other AI-focused customers, its upbeat outlook provides additional validation that Nvidia’s product roadmap and production capacity are aligned with surging data-center requirements.