Computex sends three major signals: Nvidia confirms ample TSMC capacity, AI PC chips make a joint appearance, supply and demand will remain tight in 2027.

Wall Street is closely watching the Computex conference held in Taipei, China.

On June 3, according to Chasing Wind Trading Desk, at this year’s Computex conference, Morgan Stanley’s latest research report revealed three core signals dominating the logic of technology investment for the next few years: the rise of Agentic AI, the certainty of TSMC’s production capacity, and jointly developed AI PC chips by MediaTek and Nvidia.

The most crucial assurance comes from Nvidia—which stated clearly that it has locked in sufficient TSMC capacity to support strong growth in 2027, and that the market will still be in a "supply shortage" situation at that time.

In addition, the Vera CPU, with revenue expectations as high as $20 billion, and AI PC chips expected to bring a considerable boost to earnings per share (EPS) are laying out clear profit pathways for related supply chain companies. This means that, in this non-zero-sum incremental market, hardware dividends for computing power and edge AI are just beginning.

Nvidia locks in TSMC capacity for 2027, supply-demand remains tight

The market’s greatest concern about computing power bottlenecks has been completely shattered.

Nvidia explicitly stated that it has secured enough TSMC capacity to support its strong growth in 2027. The company has successfully persuaded TSMC and other suppliers that demand will remain robust.

This releases an extremely critical expectation—that even though the supply chain is actively expanding production, by 2027, market demand is still highly likely to exceed supply.

Meanwhile, management specially emphasized that AI semiconductor market is a booming sector, and current competition is not zero-sum, nor a simple battle for market share.

Morgan Stanley believes this means the entire AI GPU supply chain (such as TSMC for manufacturing, ASE for packaging, KYEC for testing, and AllRing for equipment) will continue to benefit from this long-term supply-demand mismatch.

Nvidia and MediaTek jointly launch AI PC chips, pricing and shipment volumes revealed

Is edge AI computing necessary? Nvidia gave a practical answer. In the keynote speech, Nvidia officially released the jointly designed RTX Sparks AI PC WoA SoC (system-on-chip) with MediaTek.

Among them, MediaTek provides a customized 20-core Grace CPU for the high-end N1X chip. MSI and ASUS will officially launch devices equipped with RTX Spark (N1X) in the third quarter of 2026, perfectly matching the timeline for Microsoft’s new AI Windows system rollout in the second half of 2026.

Regarding the most concerned pricing and volume, Morgan Stanley estimates based on PC brands: AI PCs equipped with N1X chips will be priced up to $2,899, while PCs with N1 chips will be priced at $1,799.

It is estimated that in 2026, shipments of N1X/N1 SoCs will reach 5 to 8 million units. Assuming the average patent licensing fee per chip is $40, this will contribute 5% to 10% growth in MediaTek’s EPS in 2026.

At the same time, this will also directly boost the performance outlook of the WoA AI PC supply chain (such as Macronix, which provides NOR Flash).

Agentic AI triggers demand for Arm architecture, Vera CPU aims for $20 billion in revenue

Nvidia, Arm, and Qualcomm all emphasized in the keynote speech that Arm-based server CPUs will be central to supporting future Agentic AI demand.

During the investor lunch, Nvidia CEO Jensen Huang positively responded to the $20 billion revenue forecast for the Vera CPU. Management noted that this demand originates not only from AI server head nodes but also from independent CPU servers.

From the technical fundamentals, Vera’s performance reaches 1.8 times that of the highest-performing x86 CPUs, and its cores are not split between chiplets, enabling faster inter-core communication. With this absolute performance advantage, Nvidia can secure higher product pricing for Vera after deducting memory sales.

Supply chain verification shows that there are currently 2.5 to 4 million chips in stock. Additionally, the Rubin GPU has entered wafer production, with no obvious delays. Regarding the technology route, Nvidia has clearly stated to the market: it will rely on copper cable connections for as long as possible, and will only adopt CPO (co-packaged optical) technology when necessary.

~~~~~~~~~~~~~~~~~~~~~~~~

The above highlights come from Chasing Wind Trading Desk.

For more detailed interpretation, including real-time analysis and frontline research, please join [Chasing Wind Trading Desk Annual Membership]

Risk Warning and DisclaimerThe market involves risks; investment should be cautious. This article does not constitute personal investment advice, nor does it take into account individual users’ special investment goals, financial situation, or needs. Users should consider whether any opinions, viewpoints, or conclusions in this article fit their particular circumstances. Invest at your own risk.