NVIDIA's "best times" seem to have passed; now is the "AI carnival moment" for Asian tech stocks

There is no doubt that NVIDIA (NVDA.US), the "AI chip giant" with a market value of $4.2 trillion, has seen the strongest growth in its core business—data center revenue growth has significantly slowed since 2023. The H100/H200 and Blackwell architecture AI GPUs provided by this business unit offer incredibly powerful AI computing infrastructure for data centers worldwide. Additionally, significant headwinds from the Chinese market and U.S.-China regulatory risks have weakened the prospects for large-scale expansion of this business.

Meanwhile, as NVIDIA's stock price faces strong resistance from an increasingly unclear performance outlook and the overall high valuation of the U.S. stock market, it has almost stagnated. In contrast, a super bull market is unfolding for Chinese tech stocks, driven by the AI investment boom led by internet giants like Alibaba and Tencent, as well as the "domestic chip replacement" storm led by domestic chip industry leaders like Cambricon. Semiconductor giants in Taiwan and Japan, spanning AI data center computing infrastructure, consumer electronics AI chips, and advanced process chip manufacturing, are also following this bull market trajectory.

Although some long-term bullish Wall Street analysts believe that the expansion trend of AI infrastructure and high-performance network hardware business provides exciting new growth expectations, it is still much smaller in scale compared to NVIDIA's core data center business, which it heavily relies on. Unlike NVIDIA's leadership position in the AI chip field, the broad AI infrastructure and high-performance networks face intense competitive pressure, with no absolute market leader currently, remaining in fierce competition. The stock market performance also indicates that as NVIDIA's overall performance growth tends to moderate and normalize, the prospects of continuously outperforming the market are being questioned.

Some institutions are calling: NVIDIA's best days may be behind us

A recent report from investment consulting firm JR Research, based on the investment research platform Seeking Alpha, indicates that due to the normalization of growth and significant risks in regulation and execution, all investors are urged to consider carefully, as NVIDIA's best days may be behind us (Nvidia's best days look well behind us).

JR Research states that if investors have been paying attention to NVIDIA's stock performance, they would have noticed that it has given back most of its gains from the second half of the year before the earnings report in August 2025. Therefore, as we enter September, following the release of NVIDIA's FY2026 Q2 earnings report, the bullish narrative seems to have become cautious. JR Research notes that for Jensen Huang and his team, this is somewhat still "business as usual," but undoubtedly, the revenue growth of its most precious data center business has significantly slowed.

JR Research believes that although the easing of U.S.-China geopolitical relations is a constructive development, it should still be viewed with caution, especially as NVIDIA's formal re-entry into the Chinese market with the H20 AI chip has encountered substantial obstacles. The legalization of the 15% commission framework belonging to the U.S. government could be seen as a procedural step However, China seems determined to prevent H20 from easily returning to the Chinese market, citing serious concerns about the national security risks posed by potential backdoors embedded in NVIDIA chips.

The continued demand for NVIDIA's H20 chips from some Chinese companies demonstrates the strong stickiness of the CUDA ecosystem moat. However, if geopolitical regulatory barriers continue to hinder its rapid return, this moat may be further weakened by the Chinese AI ecosystem built by leaders in AI chip design such as Alibaba and Cambricon. Therefore, Wall Street analysts generally do not expect NVIDIA to fully recover its revenue from Chinese AI chips, which may also provide significant catch-up opportunities for local competitors.

In light of this, some analysts are not surprised that NVIDIA's management is attempting to downplay the recent positive impact of its recovery in China and exclude it from short-term performance guidance. NVIDIA CEO Jensen Huang stated during the earnings call that the Chinese AI chip market represents a substantial short- to mid-term growth opportunity of around $50 billion, and further delays in market entry could exacerbate NVIDIA's execution and revenue risks in the Chinese market for the remainder of FY2026.

Although Wall Street analysts have continued to raise their target price for NVIDIA over the next 12 months following the latest earnings announcement, it is also becoming increasingly clear that those moments of "surprise outperformance and significant upward guidance revisions," as well as the times when Wall Street dramatically raised NVIDIA's target price at a "doubling pace," are almost a thing of the past.

JR Research believes that it would not be surprising for Jensen Huang to once again "pull a rabbit out of a hat," trying to convince us that NVIDIA's best moments are yet to come— not only is the ramp-up of the Blackwell series AI GPU cluster progressing healthily, but the company is also betting on the trillion-dollar AI infrastructure argument, which could reshape everyone's understanding of NVIDIA's full-stack ecosystem beyond CUDA. This is also the core logic that most NVIDIA bulls adhere to in their optimism about NVIDIA reaching a $5 trillion market value.

However, the fact is that despite raising the average target price, Wall Street analysts have not changed the downward slope of their performance growth expectation curve for NVIDIA. While deepening its layout in the AI high-performance networking field is expected to enhance its diversification opportunities, the $10 billion operational scale is still relatively limited compared to its nearly $150 billion data center business.

Moreover, in the high-performance Ethernet data center business, NVIDIA must also compete head-to-head with Broadcom (AVGO.US) and Marvell (MRVL.US) in the long term— the latter two have a significant market share advantage compared to the latecomer NVIDIA and will actively defend their territory

JR Research stated that if we consider Nvidia's forward EBITDA multiple of 28 times, which is still below its 10-year average of 34.3 times, high-conviction investors might view the stock as relatively undervalued. However, if we refer to the long-term chart of NVDA above, which reflects the "wisdom of the market," questions arise: why has buying momentum failed to sustain the highs reached in August?

Interestingly, not only has buying momentum failed to continue, but most of the gains from August had also evaporated before we entered September. JR Research is not surprised, stating, "If we consider the possibility that its growth logic may continue to normalize over the next two fiscal years, I believe the market is keenly pricing in this possibility."

"Unless Jensen Huang and his undeniably outstanding team can reignite the 'Jensen magic' around the overall AI infrastructure argument over the next five years, I think expecting Nvidia to maintain excellent excess alpha returns from the current position is likely asking too much," JR Research stated.

The unstoppable surge of Asian AI computing sector stocks

In stark contrast to the slowing growth momentum of AI computing industry leaders like Nvidia in the U.S. stock market, Asian tech stocks, particularly Chinese cloud computing leaders like Alibaba, have recently experienced a meteoric rise. Driven by the explosive global demand for AI infrastructure, the computing infrastructure sector in China's A-share market and the "domestic chip substitution" related sectors amid the backdrop of U.S.-China competition have become market focal points, with multiple AI computing and chip leading stocks repeatedly hitting new highs, and their performance also significantly increasing.

With Chinese internet and cloud computing giant Alibaba reporting results that exceeded market expectations and showcasing an ambitious "super AI blueprint," the AI investment frenzy in the Chinese stock market has been further ignited, maintaining a strong upward trend for Chinese tech stocks favored by global capital this year. This super wave of artificial intelligence is enough to rival the "bull market" in U.S. tech stocks in 2023. After the earnings announcement, Alibaba's Hong Kong stock surged over 17%, with its market value soaring by over $50 billion.

Financial report data shows that Alibaba's cloud computing business revenue grew by 26% year-on-year, with revenue related to AI cloud computing maintaining triple-digit year-on-year growth for eight consecutive quarters, and capital expenditures in the second quarter increased to 38.6 billion More significantly, Alibaba has developed a new generation of AI inference chips through its subsidiary Pingtouge Semiconductor. This chip aims to fill the gap left by NVIDIA's AI GPUs in the mid-to-high-end AI chip computing power market. The chip design is compatible with NVIDIA's CUDA ecosystem and will be manufactured by domestic chip companies, although the specific chip foundry has not been disclosed. Alibaba also stated that it will proceed with its planned artificial intelligence capital expenditures and investments of up to RMB 380 billion.

Alibaba's latest performance and its exceptionally strong future AI spending outlook further bolster market sentiment towards "China's AI chip leader" and the domestic chip replacement leader, Cambricon. Recently, Cambricon's stock price and performance have resonated, highlighting the intense investment boom in domestic AI, attracting foreign institutions, including Wall Street, to flock to China's A-share and Hong Kong stock markets. In terms of performance, Cambricon's revenue for the first half of 2025 reached RMB 2.881 billion, a staggering year-on-year increase of 4347.82%, with a net profit attributable to shareholders of RMB 1.038 billion, compared to a loss of RMB 530 million in the same period last year.

Against the backdrop of optimism regarding the expansion of AI capital expenditures by Chinese cloud service providers and the surge in domestic AI chip demand, Wall Street financial giant Goldman Sachs has raised its target price for Cambricon just a week after its last adjustment. In its latest report released on September 1, Goldman Sachs raised Cambricon's 12-month target price from RMB 1,835 to RMB 2,104, an increase of 14.7%, while maintaining a "Buy" rating. The latest target price implies a 41% upside potential compared to the closing price on August 29.

This adjustment follows Cambricon's exceptionally strong performance in the second quarter of 2025. Goldman Sachs' bullish report points out that China's cloud service giants are accelerating investments in AI infrastructure, combined with supportive industrial policies from government agencies, which are collectively driving demand for domestic AI chips. As the leader in "domestic chip replacement" and a key player in the AI computing power industry chain, Cambricon stands to benefit from this trend. Additionally, revenue growth and improved operational efficiency will also help enhance Cambricon's operating cost ratio.

In the semiconductor equipment sector, the Chinese chip industry chain has long faced severe constraints from the U.S. in advanced manufacturing. In recent years, U.S. sanctions against China's chip industry chain have escalated, focusing on semiconductor equipment, raw materials, and chip manufacturing. Therefore, to achieve comprehensive domestic production in the chip manufacturing sector, various high-end semiconductor equipment required for chip manufacturing, which is essentially in the "from 0 to 1" initial development stage, has become the core focus of funding from all levels of government and private capital.

Recently, the stock prices in this sector have surged significantly, benefiting from an unprecedented "domestic chip replacement" storm in the Chinese stock market. Notably, the recent removal of Samsung, Intel, and SK Hynix's companies in China from the "verified end-user" authorization list by the U.S. means that these three chip giants may no longer be able to use any manufacturing patents, semiconductor equipment, or semiconductor raw materials based on U.S. technology in the future. This effectively vacates the market share of Samsung, Intel, and Hynix in China, accelerating the favorable conditions for domestic replacement, especially in the process of semiconductor equipment domestic replacement In addition, semiconductor equipment is also a beneficiary sector under the global trend of companies laying out AI. Currently, the demand for AI chips worldwide is extremely strong, and this explosive demand is expected to continue until 2027. Therefore, chip manufacturers such as TSMC, Samsung, and Intel will fully expand their production capacity. Coupled with storage giants like SK Hynix and Micron expanding HBM capacity, there is a need for large-scale procurement of semiconductor equipment required for chip manufacturing and advanced packaging, and even some core equipment needs to be updated. After all, AI chips have higher logic density, more complex circuit designs, and higher power and precision requirements for equipment, which may lead to higher technical requirements in processes such as lithography, etching, thin film deposition, multilayer interconnection, and thermal management, thus requiring customized manufacturing and testing equipment to meet these demands.

Therefore, semiconductor equipment giants hold the "lifeline of chip manufacturing." Currently, the Chinese chip industry chain urgently needs domestic semiconductor manufacturing equipment in the most important chip manufacturing processes such as Atomic Layer Deposition (ALD), Chemical Vapor Deposition (CVD), Physical Vapor Deposition (PVD), Rapid Thermal Processing (RTP), Chemical Mechanical Polishing (CMP), wafer etching, and ion implantation, as well as in the two advanced packaging processes of wafer Hybrid Bonding and Through Silicon Via (TSV), to accelerate the process of domestic substitution.

Oliver Cox, the Asia-Pacific equity fund manager at JP Morgan Asset Management, stated that Chinese semiconductor equipment manufacturers have the "shovel seller" advantage—regardless of how the competitive landscape of downstream chip manufacturers changes, equipment demand will continue to benefit from industrial upgrades and the wave of domestic chip substitution under the US-China rivalry. This fund manager, who manages $2.1 billion in assets, has seen his fund's performance exceed 95% of its peers this year.

Semiconductor giants in Taiwan, Japan, and South Korea, which span AI data center computing infrastructure, consumer electronics-side AI chips, and advanced process chip manufacturing, are also following this bull market trajectory. The market believes that semiconductor giants such as TSMC, Tokyo Electron, Advantest, SK Hynix, and Samsung will not only benefit from the AI infrastructure boom but will also benefit from the upcoming new round of chip industry prosperity cycle covering AI smart glasses, smartphones, and PCs, showing a strong expansion trend. In contrast, US fabless manufacturers like NVIDIA do not possess the growth attributes that span multiple product terminals and chip manufacturing.

TSMC's stock price has surged significantly this year, and in July 2025, its market value on the Taiwan stock market first surpassed the $1 trillion milestone. Since the annual low in April, TSMC's stock price on the Taiwan stock market and its ADR stock price in the US have cumulatively risen over 50%, with its market value surpassing that of Berkshire Hathaway, led by Warren Buffett, ranking ninth in the global market value rankings. This unprecedented market value reflects investors' high confidence in TSMC's prospects in the AI wave: as a core chip foundry for giants like Apple, NVIDIA, and AMD, TSMC is regarded as an indispensable "chip manufacturing cornerstone" in the global AI computing industry chain and the trend of edge AI