The 2U HGX Rubin NVL8 system provides the densest and most flexible HGX platform; and the first HGX platform to offer greater flexibility in CPU selections including @nvidia Vera CPUs alongside next-generation @AMD and @intel x86 processors. Join Alok at @NVIDIAGTC Berlin as he gives a live presentation from the booth. Learn more: #GTCBerlin# #NVIDIAGTC# #NVL8# #HGX#

$NVDA $SKHY $MU $005930 - $NVDA reduced Rubin HBM4 bandwidth from 22 to 19.2 TB/s, down 12.7%. - This implies a reduction from approximately 10.75 to 9.375 Gbps/pin. - HGX Rubin NVL8 memory capacity fell from 2.3 to 2.0 TB, down 13%. - NVL8 aggregate HBM4 bandwidth dropped from 176 to 130 TB/s, down 26%. - TrendForce previously reported that HBM supply constraints and HBM4e qualification uncertainty were prompting NVIDIA to reconsider Rubin Ultra memory configurations. > The critical issue may not be whether HBM4 can reach 10.75 Gbps, but whether suppliers can deliver sufficient volume at acceptable yields. > The NVL8 capacity reduction strengthens the possibility of HBM supply or yield constraints beyond simple clock optimization. > A seven-stack configuration remains speculative, and the bottleneck could lie in the HBM stacks or GPU memory controllers. > The 16.7% NVLink bandwidth reduction suggests HBM alone may not explain every specification change. > For $005930, $000660, and $MU, production yield and qualified supply volume at Rubin's required speeds may matter more than maximum advertised bandwidth. > Lower qualification requirements could also reshape competition among HBM4 suppliers.

Nvidia $NVDA and Oracle $ORCL are partnering to expand Vera Rubin NVL72 clusters on OCI to “multi-gigawatt scale”, with Oracle also introducing HGX Rubin NVL8 offerings and separate Vera CPU instances for agentic AI.

☀️Welcome to the Supermicro Booth at @ @NVIDIAGTC Berlin 2026! Join Juan as he gives a preview tour of our latest AI data center innovations, including our Supermicro @nvidia HGX™ Rubin NVL8 and Supermicro NVIDIA Vera Rubin NVL72. Watch the full video here:

$CSCO, $SMCI EXPAND NVIDIA AI FACTORY STACK Cisco is partnering with Supermicro to add rack-scale AI systems to its Secure AI Factory with NVIDIA portfolio, with availability starting October 2026. The offering will support $NVDA Vera Rubin NVL72 and HGX Rubin NVL8, combining Supermicro liquid- and air-cooled servers with Cisco networking. Cisco is targeting enterprise, neocloud and sovereign-cloud deployments, including NVIDIA Cloud Partner-compliant infrastructure and new validation services for large-scale AI builds.

$NVDA and $ORCL Plan Multi-Gigawatt Vera Rubin Expansion as Vera CPUs Go Live on OCI Vera Rubin NVL72 systems and standalone Vera CPUs are now generally available on Oracle Cloud Infrastructure. Each NVL72 rack connects 72 GPUs through NVLink for scale-up communication. Across racks, Oracle Acceleron integrates Spectrum-X Ethernet, Spectrum-6 switches and BlueField-4 DPUs, supporting clusters spanning hundreds of thousands of GPUs. The multi-gigawatt expansion remains planned. The CPU deployment targets agent workloads between model calls: tool execution, code testing and data retrieval. Vera combines high single-thread performance with high memory bandwidth, while OCI Kubernetes Engine lets customers scale CPU and GPU resources independently. The aim is to reduce CPU bottlenecks that leave GPUs waiting. OpenAI says it is bringing Vera into production on OCI for agentic sandboxes. OCI also plans an x86-based HGX Rubin NVL8 configuration: eight NVLink-connected Rubin GPUs, two 64-core $INTC Xeon 6 CPUs, 4TB DDR5 and 6.4Tbps RDMA networking. This provides a smaller Rubin deployment option while preserving x86 application compatibility. Source:

OpenAI is bringing NVIDIA’s Vera CPUs into production on Oracle’s cloud to run code and tools for AI agents. Initial tests showed 51% higher Python throughput and 2.2x memory bandwidth per thread vs. a leading x86 CPU Oracle plans an initial deployment of 150,000 Vera CPUs. OCI’s Vera offering includes standalone bare-metal and flexible VM instances, allowing customers to scale CPU and GPU capacity independently. NVIDIA and Oracle also announced general availability of Vera Rubin NVL72 systems on OCI, with plans to expand Vera Rubin superclusters to multi-gigawatt scale. Oracle also plans to introduce HGX Rubin NVL8 systems, offering eight Rubin GPUs in an x86-based server configuration.

AI SILICON DAILY DIGEST, TUESDAY, OCTOBER 27, 2026 DESIGN WINS • NVIDIA Vera Rubin NVL72 is now generally available on Oracle Cloud Infrastructure, per the NVIDIA and Oracle announcement of October 27. OCI Superclusters are planned to expand with Vera Rubin NVL72 systems to multi-gigawatt scale. • NVIDIA Vera CPUs are also generally available on OCI under the same announcement. • A planned HGX Rubin NVL8 offering was announced in the same release: eight Rubin GPUs, two 64-core Intel Xeon 6 CPUs, 4TB DDR5, and 6.4 Tbps RDMA. BACKFILL GAPS • SambaNova SN50 to Oracle OCI and Google TPU 8t and 8i to Oracle OCI were confirmed in official releases today, but neither product has a row in the database yet (both predate it), so the wins are documented as backfill gaps rather than recorded design wins.