Vera Rubin NVL72
AI overview
Sign in and the AI will write an overview from our coverage.
Headlines · 9
- NVIDIA Revises Vera Rubin Specifications for Volume Production
NVIDIA said the revised Vera Rubin specifications represent what its systems sustain after qualification and volume production, rather than the preliminary architectural targets published during development. The changes lower the per-Rubin GPU HBM4 bandwidth from 22 TB/s to 19.2 TB/s and NVLink 6 bandwidth from 3.6 TB/s to 3 TB/s across the platform. NVIDIA also said Vera Rubin NVL72 is shipping and that published third-party agentic benchmarks used the production specifications, but the supplied reports do not independently verify those claims.
Wccftech · 🔥 0 - Analyst projects SpaceX to receive 20,000 NVIDIA Vera Rubin racks
GF Securities analyst Jeff Pu expects SpaceX to receive 20,000 NVIDIA Vera Rubin NVL72 racks in 2027, making it the largest single buyer in a projected 96,000-rack market. Pu says the six major neoclouds are expected to buy 34,000 racks, or 35% of projected demand, and estimates Foxconn could handle up to 45% of SpaceX's data-center orders.
Wccftech · 🔥 0 - NVIDIA Revises Vera Rubin Bandwidth Specifications
IT之家 and Wccftech reported that comparisons between archived and current NVIDIA pages show lower HBM4 and sixth-generation NVLink bandwidth figures for several Vera Rubin configurations, alongside higher scale-out networking bandwidth. For the Rubin GPU, the published HBM4 bandwidth changed from a maximum of 22 TB/s to 19.2 TB/s, while NVLink 6 bandwidth fell from 3.6 TB/s to 3 TB/s. The changes also cover the Vera Rubin Superchip, NVL72 and HGX Rubin NVL8 systems. NVIDIA has not publicly explained the revision, and some product and regional pages still show older specifications. These are published specification changes rather than measured benchmark results, and the earlier 22 TB/s figure was a maximum, so the revision does not necessarily represent an equivalent decline in real-world performance across every configuration.
Wccftech · 🔥 10 - Google Cloud and Nvidia say A5X instances with Vera Rubin NVL72 open next month
Google Cloud and Nvidia announced on October 21 that A5X instances running Nvidia's Vera Rubin NVL72 system will open next month, with plans to extend the offering to the European market, according to a 36Kr brief citing Jiemian. The two companies position the infrastructure as compute for AI agent inference and reinforcement learning workloads. The report did not disclose pricing, specific regions or configuration details.
36氪快讯 · 🔥 0 - NVIDIA Introduces DSX POD Modular AI-Factory Design
At GTC Berlin on October 21, NVIDIA introduced DSX POD, a reference design for partner-built Vera Rubin AI factories housed in modular containers and integrated and validated before delivery. NVIDIA says that, when power, cooling, networking and other supporting infrastructure are already prepared at the site, deployment can be reduced from six months to as little as 14 days. The company says this could help customers begin generating production tokens and earning revenue sooner, while the earliest partner deliveries are planned worldwide for early 2027. Partners include Anduril, Armada, Bull, Crusoe, DayOne, Dell Technologies, Delta, Eaton, Foxconn, GIGABYTE, HPE, Lenovo, Pegatron, Quanta Cloud Technology, Schneider Electric, Vertiv and Wistron. NVIDIA also introduced DSX OS, an open-source, modular developer preview of internal data-center operations software.
NVIDIA Newsroom · 🔥 0 - Meta details Prometheus, an Ohio supercluster heading toward 1 million GPUs
Meta published engineering details of its Prometheus supercluster in New Albany, Ohio, saying it already runs hundreds of thousands of Nvidia GPUs across several buildings and is planned to grow to roughly 1 million operating in lockstep. The first gigawatt-scale phase is finished and was used to train Meta's Muse family of models, including Muse Wonder, on Nvidia GB200 NVL72 and GB300 NVL72 racks, with Vera Rubin NVL72 planned for a third platform. Meta says a year of engineering cut training interruptions by 99%, from nearly 1,000 per 8,000 GPUs per day in mid-2025 to about 11, and that each 72-GPU rack trains on a 64-GPU slice with eight GPUs kept as spares.
The Next Web · 🔥 0 - Musk: SpaceX-Nvidia Star Mind AI satellite computer to enter volume production next year
Elon Musk said in an Oct. 15 post on X that the Star Mind AI computer co-engineered by SpaceX and Nvidia is "far better than almost anyone expects" and will reach volume production next year, while the TechNews article body dates volume production to 2027. According to information Nvidia previously published, SpaceX is advancing the Star Mind satellite computing system using a space-optimized Vera Rubin NVL72 architecture combining Vera CPUs and Rubin GPUs to extend high-performance AI computing from ground data centers into orbit. Citing foreign media, the report says the first custom satellite built for the space environment is expected to launch in the fourth quarter of 2027, with scaled operation hoped for in 2028, though the technology remains in development and actual specifications and commercial benefits are unconfirmed.
科技新报 · 🔥 0 - ASUS launches XA P2N-E2 MGX server with dual RTX PRO 6000 Blackwell GPUs
ASUS announced the XA P2N-E2, a 2U server built on NVIDIA's MGX architecture, at the 2026 OCP Global Summit. It pairs a Vera CPU with up to two RTX PRO 6000 Blackwell server-edition GPUs and is pitched by ASUS for agentic AI and reinforcement-learning workloads, including code execution, tool calling and sandboxed compute. ASUS also showed other systems, among them the Vera Rubin NVL72-based XA VR721-E3, the air-cooled-CPU/liquid-cooled-GPU XR R1I12B-M1BHFS, the dual EPYC 9006 1U RS700A-E14B-R16U and the AMD-platform air-cooled AI system ESC8000A-E13P.
IT之家 AI · 🔥 0 - vLLM adds day-0 support for NVIDIA Vera Rubin NVL72, claims 7.8x throughput over GB200
vLLM says it now runs on NVIDIA's Vera Rubin NVL72, with nightly CUDA 13.4 and PyTorch 2.15 container images and support for models from DeepSeek, Moonshot AI's Kimi, Z.ai's GLM and MiniMax. Working with NVIDIA, Red Hat and Inferact, the project reports early results of 7.8x throughput per GPU versus GB200 NVL72 on AgentX at matched interactivity and up to 3.7x higher VLM throughput versus GB300 NVL72 in MLPerf; these are its own early figures. MiniMax separately said its M3 model is now running on Rubin with vLLM, crediting the same collaborators.
vLLM Blog · 🔥 0
Experience and discussion from the community
Share my Vera Rubin NVL72 experienceAsk about Vera Rubin NVL72
Nobody has shared their experience with Vera Rubin NVL72 yet.