Four Taiwan AI Server Makers Post Record May Revenue as Nvidia Grace Blackwell Servers Ship in Volume
Four of Taiwan's biggest AI server manufacturers posted record May revenue, according to a media report, as Nvidia Grace Blackwell-based servers shipped in volume and Vera Rubin server production ramped up. Quanta reported NT$311.5 billion, up 94.4% from a year earlier, Wistron NT$290.2 billion ($9.2 billion), up 39.2%, Wiwynn NT$84.1 billion, up 18.2%, and Inventec NT$82.8 billion, up 35.3%.
The results add supply-chain evidence to Nvidia's earlier disclosure that it had doubled production capacity as Grace Blackwell output ramps and Vera Rubin-based products begin mass production. Quanta said it expects double-digit second-quarter revenue growth from the first quarter, triple-digit growth for the full year and visibility to 2027, while Wistron forecast strong AI server growth, 20% growth in general-purpose servers and quadruple growth in networking this year.
From the sources (19 posts)
@firstadopter"Our next big training run in the fall will be done on Vera Rubins. We're really excited about that." - @OpenAI CFO Sarah Friar via @theallinpod $NVDA
@stocksavvyshayOpenAI CFO Sarah Friar says its next major training run this fall will be on $NVDA Vera Rubin platform. She also said compute remains sold out through much of 2027 with power, land, chips and talent all becoming bottlenecks
@financialjuiceNvidia CEO Jensen Huang: Memory shortage to continue for several years
@stockmktnewzNvidia CEO Jensen Huang just said that he thinks the memory shortage will last for several more years
@financialjuiceNvidia CEO Huang, SK Chairman Chey to reveal cooperation plan Monday - Newsis
@firstsquawkNvidia CEO Huang, SK Chairman Chey to reveal cooperation plan Monday - Newsis
@jukan05Jensen Huang: “Demand is enormous. From wafers to silicon photonics and cable connectors, everything across the entire industry supply chain is in short supply.”
@stocksavvyshay$NVDA CEO Jensen Huang says he expects the memory shortage to last several more years. HBM uses far more wafer capacity than regular DRAM and depends on advanced packaging so even $MU, Samsung and SK Hynix cannot fix supply just by adding
@zephyr_z9RT @jukan05: Jensen Huang: “Demand is enormous. From wafers to silicon photonics and cable connectors, everything across the entire industr…
@cnbcNvidia, SK to detail cooperation plan as Huang flags prolonged chip shortage
@stocksavvyshay$NVDA CEO Jensen Huang says demand remains enormous with shortages spanning everything from wafers to silicon photonics and cable connectors. He also teased that Nvidia may have additional announcements coming tomorrow.
@businessNvidia's Jensen Huang said the company's new Vera central processing units will use SK Hynix memory chips, as the two companies prepare to do more business in the coming year
@reutersNvidia, SK to detail cooperation plan as Huang flags prolonged chip shortage
@reutersNvidia, SK to detail cooperation plan as Huang flags prolonged chip shortage
@dnystedtNvidia CEO Jensen Huang said the 2nd half of this year will be extremely busy, and Nvidia has already doubled its production capacity in preparation, media report. He said Grace Blackwell-based production continues to ramp up now, while Ver
@semiconalphaNVIDIA @ BofA Global Technology Conference 2026 Accelerated Vera Rubin Timeline: NVIDIA confirmed its next-generation Vera Rubin architecture is in full production and scheduled to stand up ahead of schedule in Q3 2026 Standalone Vera CPU
@dnystedtFour of Taiwan’s biggest AI server makers all posted record May revenue as Nvidia Grace Blackwell-based servers ship in volume and Vera Rubin-server production ramps up, media report: Quanta NT$311.5 billion, +94.4% year-on-year Wistron NT$
@dnystedt2/5 Quanta is not only benefitting from Nvidia-based server sales, it is also now doing ODM/OEM (design and assembly) for four of the big cloud service providers and their own custom ASIC chips. The company expects double digit % revenue gr
@dnystedt3/5 Wistron expects strong growth in its AI server business, 20% year-on-year growth in general-purpose servers and quadruple % growth in networking this year.