To The Moon

All eyes on Wall Street are fixed on the Computex event taking place in Taipei.

According to recent analysis, this year's conference has highlighted three core signals that will dominate technology investment logic in the coming years: the rise of Agentic AI, the certainty of Taiwan Semiconductor Manufacturing (NYSE: TSM) production capacity, and the collaborative AI PC chip from MediaTek Inc. and NVIDIA (NASDAQ: NVDA).

The most crucial reassurance comes from NVIDIA, which has explicitly confirmed it has secured sufficient Taiwan Semiconductor Manufacturing capacity to support its robust growth through 2027, with the market expected to remain in a state of supply shortage at that time.

Key Signals from the Conference

Furthermore, the Vera CPU with its projected $20 billion revenue potential and the AI PC chips anticipated to deliver significant earnings per share accretion are carving out clear profitability pathways for related supply chain companies. This indicates that in this non-zero-sum, growth-oriented market, the hardware dividends from computing power and edge AI are only just beginning.

NVIDIA Secures Long-Term TSMC Supply

The market's greatest concern regarding a computing power bottleneck has been decisively addressed.

NVIDIA has clearly stated it has secured ample capacity from Taiwan Semiconductor Manufacturing to underpin its strong growth in 2027. The company has successfully convinced TSMC and other suppliers of sustained, robust demand.

This sets a critical expectation: despite active supply chain expansion, market demand in 2027 will likely continue to outpace supply.

Simultaneously, management emphasized that the AI semiconductor market is a rapidly expanding pie, where current competition is not a zero-sum game nor a simple battle for market share.

This analysis suggests that the entire AI GPU supply chain—including foundry partner Taiwan Semiconductor Manufacturing, packaging provider ASE Technology Holding, testing specialist KYEC, and equipment maker AllRing Tech—will continue to benefit from this prolonged supply-demand imbalance.

AI PC Chip Launch with MediaTek

Is edge AI computing necessary? NVIDIA has answered with action. During a keynote, the company formally launched the RTX Sparks AI PC WoA SoC, co-designed with MediaTek Inc..

MediaTek Inc. is providing a 20-core custom Grace CPU for the high-end N1X chip. MSI and ASUS are scheduled to launch devices featuring the RTX Spark (N1X) in the third quarter of 2026, aligning perfectly with Microsoft's (NASDAQ: MSFT) new AI Windows system launch expected in the second half of 2026.

Regarding crucial pricing and volume expectations, checks with PC brands indicate AI PCs equipped with the N1X chip may carry a price tag as high as $2,899, while those with the N1 model are expected to be priced around $1,799.

Shipments for the N1X/N1 SoC in 2026 are projected to reach 5 to 8 million units. Assuming an average royalty fee of $40 per chip, this could contribute a 5% to 10% increase to MediaTek Inc.'s 2026 earnings per share.

This development is also expected to directly boost the performance outlook for the WoA AI PC supply chain, including companies like NOR Flash provider Macronix.

Agentic AI Drives Arm Demand, Vera CPU Targets Major Revenue

In keynotes, NVIDIA, Arm, and Qualcomm unanimously emphasized that Arm-based server CPUs will be central to supporting future Agentic AI demand.

At an investor luncheon, NVIDIA CEO Jensen Huang responded positively to the ambitious $20 billion revenue projection for the Vera CPU. Management noted that this demand stems not only from AI server head nodes but also from standalone CPU servers.

From a technical standpoint, Vera offers performance up to 1.8 times that of the highest-performing x86 CPUs. Its cores are not split across chiplets, enabling faster inter-core connectivity. Leveraging this performance advantage, NVIDIA can command higher product pricing for Vera, excluding memory sales.

Supply chain checks indicate preparations for 2.5 to 4 million units are already underway. Additionally, the Rubin GPU has entered the wafer fabrication stage with no significant delays reported. On the technology roadmap, NVIDIA has clearly communicated its intention to rely on copper connections for as long as possible, adopting co-packaged optics technology only when necessary.