At the annual SIGGRAPH conference in Vancouver, Canada, various tech giants including Cisco, Dell Technologies, HPE, Lenovo, and Supermicro will incorporate the GPU into their popular 2U enterprise server systems.
Under the Hood
The 2U Nvidia RTX Pro Servers are designed to fit in standard enterprise data center racks, providing x86-based, air-cooled, PCIe-based hardware that meets enterprise IT requirements. According to Nvidia’s vice president of omniverse and simulation technologies, Rev Lebaredian, “Nvidia RTX Pro Server makes the transition from general-purpose clusters to an AI factory seamless without a complete data center overhaul.”
Initially unveiled at the Computex show in Taiwan in May, the latest RTX Pro servers powered by the RTX Pro 6000 Blackwell Server Edition GPU were offered in 4U systems by server vendors. However, the recent announcement extends the lineup to 2U form factors.
‘Blackwell for the Masses’
Analyst Matt Kimball from Moor Insights & Strategy lauded the new Blackwell GPU-powered Nvidia RTX Pro Servers as a solid choice for traditional enterprises looking to engage in AI training, tuning, or inferencing on-site. These servers also support virtual desktop infrastructure, data analytics, and other accelerated workloads.
Kimball highlighted Nvidia’s focus on enabling traditional enterprises to embrace AI practically, referring to the RTX as a perfectly sized GPU for mid-sized enterprise customers seeking to enhance manufacturing and supply chain efficiency.
Moreover, Nvidia’s strategy to equip enterprises with hardware for accelerated workloads positions them ahead of the curve in the commercial enterprise sector, according to Kimball. This move gives Nvidia a competitive edge over AMD, as the latter lacks a counterpart to the RTX Pro 6000 Blackwell Server Edition GPU.
Nvidia RTX Pro Server Performance
Nvidia executives emphasized that as enterprises acquire millions of servers annually for business workloads, they can upgrade their CPU-only systems with mainstream GPU-powered 2U servers. The new 2U Nvidia RTX Pro Servers deliver significantly better performance for traditional CPU-based workloads like data analytics, simulation, video processing, and graphics rendering. This translates to higher energy efficiency and lower total cost of ownership compared to CPU-only 2U systems.
The upcoming RTX Pro Servers, supporting Nvidia AI Enterprise and Omniverse software, offer a potent, scalable platform for AI workloads, including agentic and physical AI use cases. Scheduled for release this fall, these servers accelerate LLM inferencing six times faster than enterprise servers running Nvidia’s previous generation L40S GPUs.
Nvidia’s Additional SIGGRAPH Announcements
During the SIGGRAPH conference, Nvidia also introduced:
– New additions to the Nemotron model family for improved AI agent reasoning.
– Availability of new Omniverse SDKs and libraries for industrial AI and robotics simulation applications.
– Introduction of Cosmos World Foundation Models for faster synthetic world generation and physical AI reasoning.
– Release of new Nvidia Blackwell desktop GPUs for workstations.
This article sheds light on Nvidia’s groundbreaking efforts to revolutionize the integration of AI and accelerated workloads into traditional enterprise settings, paving the way for enhanced operational efficiency and technological advancements.
