Nvidia's Huang Urges SK Hynix to 'Please Make More' as Vera Rubin Hits Full Production

Vera Rubin Reaches Full Production

Nvidia CEO Jensen Huang used his keynote at Computex 2026 in Taipei, doubling as GTC Taipei, to confirm that the company's next-generation Vera Rubin platform has reached full production, with shipments expected to begin this autumn. Huang called it potentially the largest product launch in the history of Taiwan. The platform pairs an 88-core Vera CPU, built on Nvidia's Olympus core architecture, with a Rubin GPU, both manufactured on TSMC's 3nm process.

The five-rack system encompasses Vera Rubin NVL72 systems, Groq 3 LPX inference trays, Spectrum-6 Ethernet racks and BlueField-4 storage. Huang noted that the supply chain created for Vera Rubin is twice as large as that for the previous Grace Blackwell generation, with more than 350 factories across 30 countries involved and 150 ecosystem partners in Taiwan alone.

'Please Make More'

On 2 June, Huang visited the SK Hynix booth at the Nangang Exhibition Centre alongside SK Group Chairman Chey Tae-won, where the memory maker publicly displayed HBM4E wafers and chipsets for the first time. Huang signed the wafer with the message "Please Make More" and wrote "LOVE SOCAMM" on a next-generation server memory module. It was the second consecutive day the two executives had met, following Huang's keynote in which he named SK Hynix, Samsung Electronics and Micron as HBM4 suppliers for Nvidia's next-generation AI platform.

Industry analysts estimate SK Hynix holds the largest share of HBM4 allocation at roughly 60 to 70 percent, with Samsung at around 25 to 30 percent and Micron providing the remainder. Huang remarked that SK Hynix became a $1 trillion market cap company 70 years after its founding, a nod to the memory maker's rise amid surging AI chip demand.

A $150 Billion Bet on Taiwan

Huang reiterated Nvidia's plan to increase annual spending with Taiwanese suppliers to $150 billion, up from $100 billion currently. The figure underscores Taiwan's centrality to the AI hardware ecosystem, with TSMC, Foxconn and other manufacturers forming the backbone of Nvidia's production capacity. Chey told reporters that SK Hynix hopes to maintain its status as a major HBM supplier, adding that it is important to provide what customers want.

Pushing Into the PC

Huang also introduced RTX Spark, an Arm-based Windows PC chip developed with MediaTek that combines a 20-core Grace CPU and a Blackwell GPU, marking Nvidia's most direct push into the consumer PC market. While the company says it has secured supply for very robust growth, Huang acknowledged that supply limitations across the industry have not been fully resolved, with demand for AI semiconductors continuing to outstrip available capacity.