Stock Markets August 31, 2026 10:16 AM

Nvidia resurrects Rubin CPX hardware program, schedules production for 1Q27

Industry checks indicate revived CPX design with revised memory, power and rack architecture to operate alongside Rubin NVL72

By Sofia Navarro
Share
Twitter Reddit Facebook LinkedIn
NVDA

Industry research from analyst Ming-Chi Kuo indicates Nvidia has restarted development of the Rubin CPX product line, with mass production slated to begin in the first quarter of 2027. The refreshed CPX design alters GPU specifications, memory capacity, rack layout and interconnect topology and is intended to operate in a 1:1 pairing with the Vera Rubin NVL72 for decode workloads.

Nvidia resurrects Rubin CPX hardware program, schedules production for 1Q27
NVDA
Summarize with
ChatGPT Perplexity Claude Grok Gemini

Key Points

  • Nvidia has restarted the Rubin CPX program with production planned for 1Q27, per analyst Ming-Chi Kuo.
  • The redesigned CPX uses 168 GB HBM4, a 2,300 watt power cap per GPU, separate MGX ETL racks, and configurable systems of 64-256 GPUs organized in 64-GPU modules.
  • CPX must operate in a 1:1 pairing with Vera Rubin NVL72, handling prefill and KV cache creation that are then passed to Rubin via Ethernet RDMA; this affects AI inference infrastructure and data center deployments.

Nvidia Corp has reportedly brought its Rubin CPX program back onto its roadmap after the market had largely assumed the product was dropped. Industry checks conducted by Ming-Chi Kuo of TF International Securities show production for the revised CPX is set to begin in the first quarter of 2027.

According to Kuo, the updated Rubin CPX departs from prior CPX iterations in several technical areas. Each CPX GPU in the new design offers compute performance comparable to the Rubin GPU and carries a peak power rating of 2,300 watts. Memory is configured as 168 GB of HBM4 per CPX unit, while the Rubin GPU supports 288 GB of HBM4 and the earlier CPX design used 128 GB of GDDR7.

The physical and system architecture has also been revised. Rather than sharing rack space with Rubin, the new CPX units will be deployed in a separate MGX ETL rack. Customers will be able to choose systems populated with 64, 128, 192, or 256 CPX GPUs. The design organizes every 64 GPUs into a module composed of eight compute trays - each tray containing eight GPUs - plus a single switch tray per module.

For intra-tray scaling, NVLink will interconnect eight CPX GPUs within a tray, delivering between 1 and 1.5 TB per second of NVLink bandwidth per CPX. By comparison, each Rubin GPU offers 3.6 TB per second of NVLink bandwidth. Tray-to-tray communications will rely on Spectrum-6 Ethernet using copper links, while cross-module connections will use OSFP optical links.

Kuo emphasized an operational dependency between the CPX and Rubin systems: Nvidia requires CPX to be paired with the Vera Rubin NVL72 at a 1:1 ratio. In that workflow, CPX units perform prefill operations and construct the KV cache. The KV cache is then transferred to Rubin systems via Ethernet RDMA for decode processing.

On workload composition, Kuo noted that more than half of current AI inference workloads consist of processing input context and building KV caches. In this design context, each eight-CPX tray contains roughly 1.34 TB of HBM4 memory.


Context and implications

The revived Rubin CPX program reflects a shift in Nvidia's system-level approach, with distinct rack segmentation and a paired deployment model alongside Rubin NVL72. The revised memory and interconnect choices indicate trade-offs between capacity, bandwidth and rack-level topology that will shape how customers deploy these systems for inference workloads.

More detailed product performance and customer adoption timelines beyond the stated 1Q27 production start were not provided in Kuo's checks.

Risks

  • Dependency risk from the required 1:1 CPX-to-Rubin pairing - customers must deploy both systems together, which may complicate procurement and integration for data center operators.
  • Potential performance trade-offs due to lower NVLink bandwidth per CPX (1-1.5 TB/s) compared with Rubin GPUs (3.6 TB/s), which could affect workloads sensitive to inter-GPU bandwidth.
  • Uncertainty remains about adoption timing and broader deployment details beyond the stated 1Q27 production target, leaving questions for procurement and capacity planning in AI infrastructure and data center markets.

More from Stock Markets

Park Hotels & Resorts Climbs After BMO Upgrade, Earnings Beat, and Balance-Sheet Move Sep 1, 2026 OpenAI Rejects Apple’s Trade-Secret Accusations, Blames Company Procedures Sep 1, 2026 WPP to Axe Up to 1,000 More Roles as CEO Pushes AI-Led Restructuring Sep 1, 2026 Asian equities slip as Fed uncertainty and oil rally weigh; South Korea exports and China PMI grab attention Aug 31, 2026 Kyowa Kirin Shares Rise After Joining U.S. Drug-Pricing Agreement Aug 31, 2026