Nvidia's Computex Double Play: A New PC Chip and a Radical Cut in GPU Needs

Jensen Huang takes the stage in Taipei this Sunday with a dual-pronged strategy that could reshape both the AI data center and the personal computer. The Nvidia chief will open the Computex conference with a keynote likely to clear up the murky picture around two major hardware initiatives: the Vera Rubin server platform and the company's first serious push into Windows laptops with an Arm-based processor. The stock, sitting roughly 10% below its mid-May peak, is looking for a spark.

The N1X Processor: Taking on Intel and AMD

The most immediate competitive threat arrives in the form of the N1X chip. Built on a 3-nanometer process at TSMC, this Arm-based high-performance processor packs 20 cores and a graphics unit on par with an RTX 5070. Nvidia plans to show off the chip at Computex, which runs from June 2 to 5. Laptop makers Dell, Lenovo, ASUS and MSI are expected to demonstrate early machines, with a market launch slated before Christmas 2026.

The ambition runs deeper than raw specs. Nvidia claims the N1X can run large language models with over 100 billion parameters locally on a laptop — a direct assault on the x86 stronghold of Intel and AMD. The company is betting that local AI workloads, not just cloud-based ones, will drive the next wave of device upgrades.

Vera Rubin: Less Silicon, Far Lower Cost

Meanwhile, Nvidia’s data-center roadmap took a leap in Taipei with the Vera Rubin platform, due for first deliveries in July 2026. The numbers are eye-popping: a single supercomputer weighs two tonnes, contains about 1.3 million components and uses eight kilometres of copper cabling. More important for hyperscalers, the combination of Rubin GPUs, Vera CPUs and NVLink-6 networking is projected to cut GPU requirements during training by 75% and reduce inference costs by up to 90%.

Should investors sell immediately? Or is it worth buying Nvidia?

That kind of efficiency gain would be transformative for cloud providers, but it raises questions about future GPU unit demand even as Nvidia’s revenue soars. Analysts also flag a risk: memory costs for the Vera generation could jump more than 400%, consuming a large chunk of the bill of materials.

Taiwan: A $150 Billion Pivot

Both initiatives are anchored in an expanding Taiwanese ecosystem. Nvidia plans to lift annual spending in the region to $150 billion, up from $100 billion. The new Taipei campus, called Constellation, will house 4,000 employees when operations begin around 2030. Construction is expected to start in June or July 2026.

The pivot is driven partly by shifting revenue patterns. Sales in China and Hong Kong tumbled 50% last quarter, while Taiwan revenue jumped more than 50%. Nvidia’s MGX partner network now includes over 20 Taiwanese firms, with newcomers Cheng Uei Precision and Catcher Technology joining stalwarts TSMC, Foxconn and Quanta. According to Huang, some partners have tripled or even decupled their revenue thanks to the AI boom.

Record Financials and a Bigger Dividend

Nvidia enters this product cycle with historic results. For the first quarter of fiscal 2027, revenue reached $81.6 billion, up 85% year-over-year. The data center segment alone contributed $75.2 billion. Adjusted earnings per share jumped 140% to $1.87. The company guided for roughly $91 billion in current-quarter revenue.

The board approved an $80 billion share buyback program and raised the quarterly dividend from $0.01 to $0.25 per share. Free cash flow stood at nearly $49 billion. Institutional appetite remains strong: Mitsubishi UFJ Asset Management increased its stake by 4.7% to about $9.93 billion.

Nvidia at a turning point? This analysis reveals what investors need to know now.

Stock: Oversold but Waiting

Shares ended last week at around €181.40, nearly 49% higher over the past twelve months but about 10% below the 52-week high of €201.05. The relative strength index of 36 signals an oversold technical condition. The yield on ten-year U.S. Treasuries hovering near 5% continues to pressure high-multiple semiconductor names.

The coming days will test whether Huang’s Taipei appearances can close that gap to the all-time high. Beyond Computex, Microsoft Build in San Francisco (June 2-3) will also focus on local AI workloads, providing another catalyst for a stock that looks cheap relative to its growth trajectory — if the product roadmap delivers on its promises.

Ad

Nvidia Stock: New Analysis - 31 May

Fresh Nvidia information released. What's the impact for investors? Our latest independent report examines recent figures and market trends.

Read our updated Nvidia analysis...

en | US67066G1040 | NVIDIAS | boerse | 69456819 |