TLDR
- Nvidia’s Vera Rubin platform could improve AI inference economics by increasing output from limited data center power.
- SemiAnalysis estimates Rubin NVL72 can generate about 39% more annual revenue per gigawatt than the strongest GB300 setup.
- Rubin could deliver about 42% more modeled profit per gigawatt on key agentic AI workloads.
- The platform achieved up to 67x higher throughput per total cost of ownership than GB300 in a 170 TPS configuration.
- Nvidia said Rubin NVL72 production is ramping, with systems running at CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud, and Nebius.
Nvidia’s Vera Rubin platform could improve the economics of AI inference as data centers face tighter power limits. A new SemiAnalysis report says Rubin can produce more revenue and profit from each gigawatt than Blackwell, a factor that could influence investor attention around NVDA stock as cloud providers expand AI capacity.
NVDA stock gains another AI efficiency angle
SemiAnalysis’ latest AgentX benchmark found that Rubin NVL72 produced about 39% more annual revenue per gigawatt than the strongest GB300 setup. The report also estimated 42% more modeled profit per gigawatt on key agentic AI workloads.
Power efficiency has become a central issue for AI data centers. Cloud companies can serve more users from the same electricity supply when systems generate more tokens for each megawatt consumed.
The report also measured Rubin’s performance against total ownership costs. In one test running at 170 tokens per second, Rubin delivered up to 67 times the total throughput per TCO of GB300.
SemiAnalysis linked the gains to Nvidia’s combined hardware and networking design. The platform joins the Rubin GPU, Vera CPU, NVLink 6, ConnectX-9, BlueField-4 and Spectrum-6 into one system that targets long-context and repeated-turn AI workloads.
Power limits shape AI infrastructure spending
The focus on power comes as data center operators face limits on available electricity in several markets. Buying more accelerators alone cannot solve that problem when facilities lack enough power to run additional systems.
Higher output from each megawatt could therefore allow operators to support more AI activity inside fixed power budgets. For NVDA stock, investors may watch whether customers adopt Rubin widely as infrastructure spending shifts toward efficiency and operating costs.
Nvidia announced Vera Rubin in March and said partners would begin offering Rubin systems in the second half of 2026. The company positioned Rubin as the next step after Blackwell for large-scale AI training and inference.
By July, Nvidia said CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure and Nebius had Rubin racks running as NVL72 production ramped. Nvidia also said its production network spanned more than 350 factory sites across 30 countries.
Those deployments give cloud operators early access to Rubin systems as Nvidia expands production during the year through partner manufacturing networks across major global markets.
Get 3 Free Stock Ebooks
Discover top-performing stocks in AI, Crypto, and Technology with expert analysis.
- Top 10 AI Stocks - Leading AI companies
- Top 10 Crypto Stocks - Blockchain leaders
- Top 10 Tech Stocks - Tech giants
