NVIDIA’s Next-Gen Rubin GPUs Enter Production with HBM4 Samples from Leading DRAM Manufacturers

NVIDIA has initiated the production of its cutting-edge Rubin GPUs, while simultaneously securing samples of the next-generation HBM4 memory from leading suppliers in the industry.

NVIDIA’s Rubin GPUs: Pioneering AI Solutions Set for 2026

During a recent appearance at GTC 2025 in Washington, NVIDIA’s CEO Jensen Huang unveiled the impressive Vera Rubin Superchip. This groundbreaking technology features two powerful GPUs stacked alongside the next-gen Vera CPU and a plethora of LPDDR memory surrounding it. The Vera Rubin Superchip is positioned to revolutionize AI computing within data centers, with promising updates regarding its production timeline.

Reports from UDN indicate that during a recent trip to Taiwan, Huang confirmed that the Rubin GPU is already making its way into production at TSMC. This significant development follows a prior statement highlighting the arrival of initial Rubin GPUs in their labs—a swift advancement from early stages to production in just a matter of days.

NVIDIA Rubin AI GPUs with HBM4 Memory from Suppliers

Regarding product and supply chain timing, Jensen Huang pointed out that demand for Blackwell is strong and not just for GPUs.“NVIDIA is also manufacturing CPUs, network chips, switches, and many other chips related to Blackwell.”He also revealed that the next-generation Rubin is already entering the production line.“We have already seen Rubin on the production line.”TSMC is working very hard to support the related demand.

via UDN

As the Rubin GPUs ramp up production, the current-generation Blackwell and Blackwell Ultra GPUs continue to experience high demand. This surge in interest has prompted TSMC to increase its 3nm production capacity by 50% in anticipation of the Rubin GPUs.

C. C.Wei, President of TSMC, noted that NVIDIA is seeking additional wafers and chips. Although the exact quantity requested remains a closely guarded secret, there is no doubt that it will be substantial, given the soaring demand for Blackwell GPUs.

Vera Rubin Superchip Specifications Highlights

Additionally, the latest reports indicate that NVIDIA has acquired samples of HBM4 memory, which is set to power the Rubin GPUs, from various manufacturers. Historically, NVIDIA has sourced DRAM from multiple suppliers to mitigate risks posed by shortages, and they are likely to continue this strategy moving forward.

NVIDIA anticipates that Rubin GPUs could enter mass production by the third quarter of 2026, or potentially even sooner. It’s crucial to distinguish between ‘production’ and ‘mass production, ’ as they refer to different levels of manufacturing readiness. The Rubin GPUs represent the next evolution in AI technology, significantly bolstered by a landmark $100 billion partnership with OpenAI, which will utilize these advanced accelerators in its data centers.

Source & Images