Analysis of the Reasons Behind the Computing Power Hardware Sector Leading Declines on the ChiNext Board and the Current Status of the AI Investment Boom

Based on the latest market data and industry trends, I will provide an in-depth analysis of the reasons behind the computing power hardware sector leading declines on the ChiNext Board and the current status of the AI investment boom.

---

According to the latest survey by TrendForce, NVIDIA adjusted the HBM4 specifications for its Rubin platform in Q3 2025, raising the Speed per Pin requirement to above 11Gbps. This forced the three major HBM suppliers (SK Hynix, Samsung, Micron) to revise their designs, pushing mass production to the end of Q1 2026[1]. This change directly impacted market expectations for high-end AI computing power chip supply, putting pressure on related concept stocks.

At the 2026 CES exhibition, Jensen Huang announced that the next-generation Rubin chip servers support liquid cooling and do not rely on traditional chilled water units. This statement caused a sharp plunge in stocks related to data center cooling equipment: Johnson Controls fell by over 11% at one point, and Modine Manufacturing dropped by as much as 21%[2]. The market is concerned that demand for traditional data center cooling equipment may shrink, triggering a re-evaluation of the entire AI infrastructure industry chain.

The investment logic for AI chips is undergoing profound changes. The launch of NVIDIA’s Rubin platform marks the evolution of AI investment from “computing power stacking” to “TCO (Total Cost of Ownership) reduction”[3]. Investors are re-examining the valuation logic of computing power hardware companies, shifting from a pure focus on computing power competition to prioritizing energy efficiency ratio, ecological synergy, and the ability to provide full-stack solutions.

In 2025, the Wind Computing Power Theme Index surged by as much as 77.5%, ranking among the top popular indices. Taking Envicool as an example, its gain exceeded 230% in 2025[4]. A large number of profit-taking positions have accumulated after the continuous rise, and some capital has chosen to lock in profits, increasing short-term adjustment pressure.

---

: The main line of the current AI industry is accelerating its shift from “parameter and ranking competition” to “capability realization and commercialization validation”[5]. This transformation means:

• Hardware Side
  : Shifting from purely pursuing peak computing power to focusing on effective computing power and cost per token
• Application Side
  : Shifting from proof of concept to commercial implementation

: Market capital is shifting from computing power hardware to AI application fields. In the morning session on January 12, bullish sentiment towards AI applications spread across the board: Yidian Tianxia secured two 20% daily limit ups, and the ChiNext AI ETF (159363) surged 3% to hit a new record high[6].

The domestic computing power industry chain continues to receive policy support. On January 7, 2026, the Ministry of Industry and Information Technology issued the “Action Plan for Integrating and Empowering Industrial Internet and Artificial Intelligence”, proposing to accelerate the integrated application of industrial Internet with general computing centers, intelligent computing centers, and supercomputing centers[7]. Meanwhile, SMIC South received an additional investment of approximately US$7 billion, marking that the expansion of advanced manufacturing processes in China is ahead of expectations, which is expected to continue to benefit the capacity release of domestic computing power chips.

Bank of China Securities pointed out that AI is entering the era of “Physical AI”, with robots and autonomous driving serving as ideal carriers[8]. Tesla’s mass production plan for the Optimus humanoid robot includes achieving mass production of 50,000 units in the first quarter of 2026, with costs reduced to below US$20,000. This trend opens up new application scenarios and market space for AI computing power.

---

Risk Factors	Opportunity Factors
HBM4 delay impacts supply chain confidence	Accelerated commercial implementation of AI applications
Need to digest valuations after large previous gains	Continuous intensification of domestic substitution policies
Market sentiment requires time to recover	Continuation of the spring rally

1. Focus on the AI Application Side
   : Application software companies with high-frequency rigid demand scenarios or clear monetization paths have the relative advantage of “light computing power, heavy scenarios”

2. Accumulate Domestic Computing Power Stocks on Dips
   : Domestic substitution targets such as Cambricon and Moore Threads are dragged down by sentiment in the short term, but the long-term logic of localization remains unchanged

3. Focus on Tracks with Marginal Improvements
   : Robotics, AI applications, non-ferrous metals, and domestic computing power are listed as key focus areas for January[3]

---

The computing power hardware sector leading declines on the ChiNext Board does not mean the AI investment boom is cooling down, but rather

. As the AI investment logic shifts from “betting on shovels” to “looking for gold mines”, the market is undergoing a style rotation from universal hardware rallies to application implementation. Investors should follow this trend and focus on AI enterprises with real commercial closed loops and implemented scenarios, rather than simply betting on computing power hardware.

[1] Wall Street News - NVIDIA Rubin Platform Delay: Precursor to AI Bubble Burst, or a Golden Opportunity? (https://shishixinwen.news/news/wallstreet/live/3032998)

[2] EET China - Jensen Huang Says No Need to Rely on Chilled Water Units, Cooling Equipment Stocks Plunged by 20% at One Point (https://www.eet-china.com/mp/a465647.html)

[3] Sina Finance - A-Share Market Expected to Continue Uptrend, Main Line of Tech + Cycle Bull Market in January Remains Unchanged (https://finance.sina.com.cn/stock/report/2026-01-11/doc-inhfxvtm9125030.shtml)

[4] Securities Times - Musk Endorses, China’s AI Computing Power May Lead the World! High-concentration Chip Concept Stocks Released (https://www.stcn.com/article/detail/3584968.html)

[5] Cailianshe - [Focus Review] ChiNext Index Weakly Fluctuates, Down 0.82%, Weighted Stocks Such as Finance Slump (https://finance.eastmoney.com/a/202601083612863850.html)

[6] Sina Fund - Bullish Sentiment on AI Applications Spreads! ChiNext AI ETF (159363) Surges 3% (https://finance.sina.com.cn/money/fund/etf/2026-01-12/doc-inhfcnu7667187.shtml)

[7] East Money - [Focus Review] ChiNext Index Weakly Fluctuates, Down 0.82% (https://finance.eastmoney.com/a/202601083612863850.html)

[8] AASTOCKS - Bank of China Securities: Entering the “Physical AI” Era, Robotics and Autonomous Driving Expected to Take Center Stage (http://www.aastocks.com/sc/usq/news/comment.aspx?source=AAFN&id=NOW.1494395&catg=4)