Technical Deep Dive
The current rotation is not a simple sector shift—it is a structural deleveraging event driven by three interconnected mechanisms: liquidity evaporation, valuation mean-reversion, and policy-driven capital reallocation.
Liquidity Evaporation Mechanism: In a shrinking-volume market (today's total turnover fell below 800 billion RMB, down 15% from the 30-day average), the bid-ask spread on small-cap, high-turnover stocks widens dramatically. Consecutive-limit-up stocks, which rely on continuous retail order flow, become the first victims. When the buying queue thins, the price support collapses. This is a classic 'crowded trade unwind'—the same algorithms and retail strategies that amplified the rally now amplify the crash. The technical trigger is the liquidity threshold: once daily turnover drops below a stock's 20-day average by 30%, forced selling by leveraged retail accounts and quantitative funds accelerates.
Valuation Mean-Reversion: The valuation spread between the top-quintile PE stocks (tech/hype) and bottom-quintile PE stocks (state-owned/industrial) has reached 4.5x, a level that historically preceded 80% of major rotation events in the past decade. The following table shows the current valuation dispersion:
| Sector | Average P/E (TTM) | Price-to-Book | Dividend Yield | 5-Year PE Percentile |
|---|---|---|---|---|
| AI Hype Stocks (limit-up) | 85x | 6.2x | 0.3% | 98th |
| State-Owned Cloud | 12x | 0.9x | 3.8% | 12th |
| Power Equipment | 15x | 1.1x | 3.2% | 18th |
| Silicon-Based Materials | 18x | 1.3x | 2.5% | 22nd |
| Broad Market (CSI 300) | 14x | 1.4x | 2.8% | 35th |
*Data Takeaway: The AI hype stocks are trading at 7x the P/E of state-owned cloud and 5.5x that of power equipment, yet with zero dividend yield. This is not a premium for growth—it is a speculative bubble that will deflate as capital rotates to sectors with real earnings and policy backing.*
Policy-Driven Reallocation: The Chinese government's recent directives on 'new productive forces' explicitly prioritize digital infrastructure (cloud, data centers) and energy security (grid upgrades, silicon materials). The State Council's 2025-2027 plan allocates 1.2 trillion RMB to these sectors, with 40% earmarked for state-owned enterprises. This creates a 'policy floor' for valuations in these sectors, making them natural safe havens during risk-off periods.
For readers interested in the algorithmic side, the open-source repository `quantlib/risk-parity` (GitHub, 2,300 stars) provides a Python library for constructing rotation-aware portfolios that dynamically adjust sector weights based on liquidity and valuation signals. Another relevant repo is `microsoft/qlib` (14,000 stars), which includes a 'rotation factor' module that backtests sector momentum and mean-reversion strategies.
Key Takeaway: The technical setup is textbook for a value rotation: extreme valuation dispersion + liquidity contraction + policy catalyst. The market is not irrational—it is repricing risk correctly.
Key Players & Case Studies
Three specific companies illustrate the dynamics at play:
1. State-Owned Cloud: China Telecom's Cloud Unit (天翼云)
China Telecom's cloud business reported 2024 revenue of 120 billion RMB, growing 65% YoY, yet the parent company trades at just 0.9x book value. This is a classic 'value trap' that is actually a 'value opportunity'—the market has not priced in the cloud division's profitability, which is now 18% net margins. The stock has been flat for 12 months while its private cloud competitors (Alibaba Cloud, Tencent Cloud) trade at 4-6x sales. The rotation into state-owned cloud is a recognition that the government will prioritize its own infrastructure for sensitive data workloads.
2. Power Equipment: NARI Technology (国电南瑞)
NARI is the dominant player in China's ultra-high-voltage (UHV) grid equipment, with 45% market share. Its P/E of 14x is near its 10-year low, despite a 22% CAGR in order backlog over three years. The company benefits directly from the 500 billion RMB State Grid investment plan for 2025-2027. Yet, the stock has been neglected because it lacks 'AI narrative' excitement. The rotation is a correction of this mispricing.
3. Silicon-Based Materials: Tongwei Co. (通威股份)
Tongwei, a polysilicon and solar cell manufacturer, has seen its stock drop 60% from 2022 highs due to overcapacity fears. However, the company now produces silicon at the lowest cost globally (under $5/kg), and the recent stabilization of polysilicon prices has restored gross margins to 25%. At 1.3x book value, it offers a 2.5% dividend yield and is a direct beneficiary of the global solar installation boom (projected 650 GW in 2025).
Competitive Comparison Table:
| Company | Sector | P/E | P/B | Dividend Yield | Revenue Growth (YoY) | Policy Exposure |
|---|---|---|---|---|---|---|
| China Telecom (Cloud) | State Cloud | 12x | 0.9x | 3.8% | 65% | High (Govt contracts) |
| NARI Technology | Power Equip | 14x | 1.1x | 3.2% | 22% | High (State Grid) |
| Tongwei Co. | Silicon Materials | 18x | 1.3x | 2.5% | 15% | Medium (Solar demand) |
| Kweichow Moutai | Consumer (benchmark) | 30x | 8.5x | 1.8% | 10% | Low |
| BYD | EV (hype) | 35x | 4.2x | 0.5% | 30% | Medium |
*Data Takeaway: The three value rotation beneficiaries all trade below 1.5x book value and offer dividend yields above 2.5%, while the 'hype' stocks (Moutai, BYD) trade at 4-8x book with negligible dividends. The market is repricing safety over growth.*
Industry Impact & Market Dynamics
This rotation reshapes the competitive landscape in three ways:
1. Capital Allocation Shift: Over the next 6-12 months, we expect a 20-30% reallocation of A-share active fund flows from technology/consumer sectors to industrial and state-owned sectors. This mirrors the 2016-2017 rotation when the 'supply-side reform' pushed capital into coal, steel, and power. The current catalyst is 'new productive forces' policy, which is more durable because it involves actual infrastructure spending (cloud, grid) rather than capacity cuts.
2. Business Model Implications: For state-owned cloud providers, the rotation means they can finally raise equity at reasonable valuations to fund expansion. For power equipment companies, it means lower cost of capital for R&D in smart grid and energy storage. For silicon material producers, it provides a buffer against further price wars.
3. Market Data & Growth Projections:
| Metric | 2024 Actual | 2025E | 2026E | Source |
|---|---|---|---|---|
| China Cloud Infrastructure Spend | $45B | $58B (+29%) | $72B (+24%) | IDC |
| State Grid Investment | $85B | $95B (+12%) | $105B (+11%) | State Grid Corp |
| Global Polysilicon Demand (tons) | 1.2M | 1.5M (+25%) | 1.8M (+20%) | BloombergNEF |
| A-Share Rotation Fund Flow (est.) | — | $120B into value | $150B into value | AINews model |
*Data Takeaway: The fundamental growth in these sectors is real and measurable. The cloud infrastructure market is growing 29% annually, yet the state-owned cloud subsegment trades at a 60% discount to private peers. This is the largest valuation gap in the sector since 2018.*
Key Insight: The rotation is not a 'defensive' move—it is a 'structural' move. Investors are not hiding; they are repositioning for the next growth cycle, which will be led by physical infrastructure (grid, cloud, materials) rather than digital speculation (AI chatbots, metaverse).
Risks, Limitations & Open Questions
1. Policy Dependency Risk: The entire value rotation thesis rests on continued government spending. If fiscal tightening occurs (e.g., due to local government debt concerns), state-owned cloud and grid investments could be delayed. The 2025 budget already shows a 5% cut in central government general expenditure, though infrastructure is protected.
2. Commodity Price Risk for Silicon Materials: Polysilicon prices could resume their decline if global solar demand disappoints (e.g., due to US tariffs or EU slowdown). Tongwei's cost advantage is real, but a 20% price drop would erase its margin buffer.
3. Liquidity Trap in Value Stocks: The very safety that attracts capital—low volatility, high liquidity—can become a trap if the broader market enters a prolonged bear phase. Value stocks can still fall 20-30% in a systemic crisis, as seen in 2022.
4. Open Question: Is the AI Hype Cycle Truly Over? The selloff in AI stocks may be a correction, not a reversal. If a new breakthrough (e.g., GPT-5 level model) emerges, capital could flood back into tech. The rotation's durability depends on whether AI monetization disappoints further.
5. Ethical Concern: The rotation favors state-owned enterprises, which may crowd out private innovation. If capital becomes too concentrated in SOEs, the broader ecosystem of private AI startups could suffer, reducing long-term dynamism.
AINews Verdict & Predictions
Verdict: The rotation is rational, overdue, and likely to persist for at least 6-9 months. The valuation gap between hype and value is the widest in a decade, and the policy tailwinds for state-owned cloud, power equipment, and silicon materials are stronger than for any other sector. The market is not 'confused'—it is finally pricing in reality.
Predictions:
1. By Q4 2025: State-owned cloud stocks will outperform the CSI 300 by 25-35%, driven by earnings upgrades and multiple expansion from 0.9x to 1.5x book value.
2. By Q1 2026: Power equipment stocks will see a 40% rally as the State Grid's 2025 investment plan is fully deployed and order backlogs convert to revenue.
3. By mid-2026: Silicon materials stocks will stabilize and begin a gradual recovery as oversupply clears and demand from global solar installations absorbs excess capacity.
4. Risk to watch: If the CSI 300 breaks below 3,200 (current: 3,450), the entire value thesis could be temporarily invalidated by systemic risk. In that case, cash is king.
What to watch next: The key signal is the weekly turnover of the top 10 limit-up stocks. When it drops below 10 billion RMB (currently 35 billion), the rotation will accelerate. Investors should set alerts for this metric and deploy capital into the three value sectors only after the panic selloff in hype stocks is complete.