Tomorrow's computing power main upgrade node attack strategy

In the previous post, I discussed the offensive strategy of betting on AI or computing power entering a main rally. However, things often don’t go as planned, and the market is actually following the third scenario in the post — metal strength, meaning abandoning the main rally based on AI or computing power. [Taogu Ba]

Tomorrow, another main rally point is expected. It will be a computing power main rally, and the basic strategy remains the same. When computing power shows support, it enters a main rally; if metals or chemicals are strong, then abandon the AI or computing power main rally.

Before discussing the conditions for a main rally, I want to introduce some concepts about computing power stocks to new friends, so they understand what these stocks are.

Yunnan Energy Holdings — holds stakes in innate computing power, Internet Data Centers (IDC).
Farsight — holds stakes in Putian Farsight Optical Communications, fiber optics.
Hua Sheng Tiancheng — a “system integrator” and “general contractor” for computing infrastructure.
Zhongtian Technology — fiber optics, the physical transmission medium connecting internal and external computing clusters. No matter how powerful GPUs are, data still needs to be transmitted via fiber optics.
Honghe Technology — electronic fabrics, the upstream of computing power. Computing chips require PCB for interconnection and signal transmission, and electronic fabrics are the key material determining PCB performance.

The conditions for entering a main rally are as follows:

  1. The highest market leader, Yunnan Energy Holdings, opens with a one-word limit-up, which can be broken but must quickly recover.
  2. The second-highest market leader, Farsight, opens with an upward indication and continues to hit limit-ups.
    The occurrence of 1 and 2 sets the mood and forms the basis for a main rally. The principle is that even the top limit-up stocks can go higher and generate profits, encouraging others to go long.
    If these conditions don’t occur, market funds will likely have the idea to push prices higher and then sell off, and those willing to go long will shift to metals and chemicals, meaning the main rally in computing power will not happen.

Therefore, tomorrow’s bullish scenario involves two opposing forces:
On one side, computing power; on the other, metals or chemicals.

First, about computing power. Where should one start? The answer is not to chase purely emotional computing power stocks, because on the first day of a holiday rally, emotional stocks often suffer heavy setbacks. Instead, focus on logical, performance-based computing power stocks — tangible, visible computing power. Metals and chemicals have been strong for a few days because their price increases are tangible.

The phrase “tangible” in the computing power sector refers to rising prices of related materials, such as computing chips, fiber optics, and electronic fabrics.

Stock selection: For account stability, avoid Yunnan Energy Holdings and Farsight, because after the recent limit-up streaks, there was a loss effect. Instead, focus on capacity stocks.
Note: The following is subjective, explaining the logic only, not investment advice.

  1. Token Going Global
    Concept logic: China’s AI large models are widely used by global developers via API services. The process involves: AI large models consuming China’s electricity through GPU computing power, transforming it into practical token services (like code generation, content analysis), which are then cross-border delivered to global users. Each token consumed signifies burning computing resources, triggering explosive and sustained demand for China’s cost-advantaged underlying computing infrastructure.
    Recognized stock:
    Hua Sheng Tiancheng.
    Stock logic: Domestic computing infrastructure, as a certified ISV partner for Huawei’s top intelligent computing, jointly building provincial and city smart computing centers, deeply tied to the wave of domestic computing.
    Logic route: Token going global / AI demand explosion → demand for smart computing center construction → Hua Sheng Tiancheng (smart computing center general contractor) → hardware procurement (servers, fiber optics, optical modules, etc.) → delivery of computing services.

  2. Fiber Optics
    Concept logic: Optical rods are the core raw material of fiber optics, occupying a high point in industry chain value and technology, accounting for about 70% of fiber optic costs. Demand for optical rods is exploding, but supply is constrained by capacity, leading to high prices and long order lead times. The core driver, like token going global, is the AI infrastructure boom. For example, to meet the low-latency, high-bandwidth interconnection needs of GPU clusters, large AI data centers consume several times more fiber optics than traditional data centers. Glass giant Corning has signed a multi-year, $6 billion deal with Meta (Facebook’s parent company) to supply fiber optic cables for AI data centers.
    Recognized stocks:
    Zhongtian Technology.
    Trend: Long Fiber Optic is the trend.
    Stock logic: Companies with fiber rod production capacity are the most direct beneficiaries, controlling high-profit segments of the industry chain and poised to benefit from both volume and price increases.

  3. Electronic Fabrics
    Concept logic: The average selling price of electronic fabrics has risen from 3.74 yuan/meter at the end of 2024 to 4.97 yuan/meter, a 33% increase. The price hike in February 2026 was even more significant, and the market is generally bullish. Historically, the high point for electronic fabrics is just over 8 yuan/meter, leaving room for further growth. The core driver, like token going global and fiber optics, is massive AI-driven demand. Due to equipment constraints and production shifts, there is a tight supply of high-end special fabrics and general electronic fabrics.
    Recognized stock:
    Honghe Technology.
    Stock logic: Companies with high-end special fabric production capacity are the main beneficiaries. Under AI influence, high-end electronic fabrics with low dielectric constant and low thermal expansion are in tight supply, and prices are expected to continue rising.

Now, about metals and chemicals.
If the main rally in computing power fails, bullish funds will shift to metals and chemicals. The seesaw relationship.

  1. Tungsten (mentioned three days ago in the previous post)
    Concept logic: In 2025, global tungsten prices are expected to rise over 220%, far outperforming gold during the same period. The tungsten ore index (884857.WI) hit a new high during trading on February 25. On the supply side: China’s control and resource depletion create rigid constraints. By 2026, overseas tungsten mine capacity (non-Chinese controlled) is unlikely to add more than 5,000 tons, unable to ease the global supply-demand gap. The Trump administration plans to use Pentagon-developed AI programs to set reference prices for key minerals, with tungsten being one of the initial focus minerals. Under the backdrop of great power competition, controlling the export of strategic materials becomes a long-term strategic tool.
    Recognized stocks:
    Jiang Tung Equipment.
    Stock logic: State-owned tungsten industry chain integration. Upstream: no own mines, relying on Jiang Tung Group for tungsten ore. Midstream and downstream: In February 2026, announced a 1.882 billion yuan private placement to acquire:

  2. Jiang Tung Hard Alloy to strengthen downstream processing.

  3. Ganzhou Huamao Tung Materials to enhance midstream smelting.

  4. Jiujiang Nonferrous Metal Smelting to develop tantalum-niobium new materials.
    Zhangyuan Tungsten.
    Stock logic: Private tungsten mining leader. Owns five mining rights including Taoxi Pit and Xin’anzi, with tungsten reserves of about 300,000 tons, and higher ore grades than industry average.

  5. Phosphorus Chemical Industry
    Concept logic: China’s phosphorus resources are limited, and tightening environmental policies create rigid supply constraints. Coupled with agricultural needs (spring fertilization, glyphosate) and new energy demands (lithium iron phosphate), and the geopolitical premium from China and the US both treating phosphorus as a strategic resource, this drives the phosphorus chemical industry into a long-term boom cycle. As of January 2026, the domestic spot price of 30% grade phosphate ore has been stable around 1,000 yuan/ton for nearly three years. The spot price of 95% glyphosate raw material remains between 23,500 and 24,000 yuan/ton. In early February, phosphate fertilizer prices increased by 2.5%.
    Recognized stocks:
    Yuntianhua (capacity stocks).
    Stock logic: Leading phosphorus chemical company, benefiting directly from rising phosphate ore prices due to high-quality phosphate sources.
    Jinzhengda (limit-up stock).
    Stock logic: Fully integrated phosphorus chemical industry chain. Upstream: Ma Lu Cao phosphate mine, achieving self-sufficiency after commissioning, reducing costs. Midstream: Processing, with traditional phosphate fertilizer and advanced phosphorus chemicals. Downstream: Industry extension, planning for lithium iron phosphate and other new energy materials.

I usually don’t write about logic because explaining it can influence psychology, making people want to buy more. This is the first time I’m doing this, so I hope friends stay objective and rational. If a main rally truly occurs, even if you buy a day late, it doesn’t matter.

View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
0/400
No comments
  • Pin

Trade Crypto Anywhere Anytime
qrCode
Scan to download Gate App
Community
English
  • 简体中文
  • English
  • Tiếng Việt
  • 繁體中文
  • Español
  • Русский
  • Français (Afrique)
  • Português (Portugal)
  • Bahasa Indonesia
  • 日本語
  • بالعربية
  • Українська
  • Português (Brasil)