To The Moon

Jensen Huang's unveiling of a new processor has triggered a significant surge across related technology stocks.

NVIDIA's (NVDA) after-hours trading saw a sharp increase of nearly 3%.

In the Taiwan market, NVIDIA's partners experienced widespread gains. Quanta Computer surged by up to 9.9%, Wistron rose 9.8%, and ASUS jumped 10%.

MediaTek's stock price also climbed over 5%.

The catalyst was NVIDIA CEO Jensen Huang's keynote at Computex in Taipei, where he introduced the new N1X processor, co-developed with Microsoft. This chip will be integrated into the new RTX Spark super chip, set to debut this fall in next-generation Windows PCs from Microsoft, Dell, HP, ASUS, Lenovo, and MSI.

Huang described the innovation as a fundamental reinvention of the personal computer, comparable to the transformation of mobile phones into smartphones. He stated that AI agents will run on all these new devices.

He added that Microsoft and NVIDIA are jointly reinventing the PC, marking the first complete redesign of the PC product line in 40 years.

An NVIDIA spokesperson indicated initial plans include over 30 laptops and 10 desktops featuring the new chip.

The inaugural PC processor combines two types of NVIDIA's flagship chips with 128GB of unified memory. It fuses the NVIDIA Blackwell GPU with a new N1X CPU based on Arm architecture and custom-designed by MediaTek.

The RTX Spark could significantly reshape the PC industry. The current AI boom is already driving major changes, with Arm-based processors like NVIDIA's gaining ground against the traditional x86 architecture championed by Intel and AMD. Huang projected this could become a $200 billion industry.

NVIDIA noted in February that CPUs were becoming a "bottleneck" amid surging AI agent workloads. In March, the company launched its Vera CPU for data centers. While GPUs excel at the parallel computations needed for training large models, CPUs provide the general-purpose computing required for data access and distribution to multiple agents.

The new PC processor will be manufactured using TSMC's 3-nanometer technology.

The market has anticipated NVIDIA's Arm-based PC chip for years. Development began in 2023 as part of Microsoft's initiative to promote Arm processors for enterprise computers. NVIDIA's spokesperson stated the company has collaborated with Microsoft on the chip for "many, many years," claiming it will be "much more powerful, much higher performance, much more efficient" than traditional x86 processors.

Intel, the pioneer of the x86 architecture in the 1970s, also announced its new Xeon 6+ data center CPU at Computex.

Recently, many companies have shifted towards the more power-efficient Arm architecture, which gained mainstream popularity with the first iPhone in 2007.

Apple now produces Arm-based processors for its computers, launching a new, higher-priced MacBook with the latest M5 chip in March. That same month, Arm released its first self-developed CPU. Reports also indicate AMD is working on an Arm-based PC chip.

The first laptops with NVIDIA's new chip will be as thin as 14mm, positioned in the premium segment, and will also be used in some compact desktops. While RTX Spark will eventually expand to different price points, NVIDIA currently targets creators, AI developers, and gamers seeking very thin, lightweight, portable laptops or compact desktops.

NVIDIA stated more performance metrics will be revealed closer to the chip's fall launch. Currently, the RTX Spark's performance is "roughly equivalent" to NVIDIA's leading RTX 5070 laptop GPU.

Huang also announced that NVIDIA's Vera data center CPU is now in full production. He stated NVIDIA is producing millions of CPUs for "a market that didn't exist before." Vera will be available starting this fall, with early customers including Anthropic, OpenAI, xAI (SpaceX), Dell, Oracle, and CoreWeave.

Huang called this a new primary growth engine, emphasizing that these CPUs must be high-performance and extremely energy-efficient to maximize deployment in AI factories without consuming power needed for token generation.

Ian Buck, NVIDIA's VP of Hyperscale and HPC, noted that high-speed CPUs are now critical for keeping AI factories running, adding that Vera currently generates tokens 1.8 times faster than x86, which will enhance overall AI agent performance and data center token revenue.

Additionally, the AI PC concept sector saw collective gains in the A-share market.