Multi-Model Consensus Report | March 2026
5-Model Analysis: Claude Opus 4.6, Claude Sonnet 4.6, GPT-5.2, GPT-5.1, Gemini 3 Pro
Framework: Universal Rules + Cyclical Rules (CRule 1–10) + Day1Global
Subject: 中钨高新 (China Tungsten High-Tech, 000657.SZ)
TL;DR — Executive Summary
| Metric |
Value |
| Stock Price |
66.2 RMB |
| Market Cap |
151B RMB (~$20.7B) |
| PE (TTM) |
~135x |
| PB |
~15x |
| 1-Year Return |
+600% |
| Consensus Rating |
SELL / TAKE PROFIT (5/5 models) |
| 12M Base Target |
27–58 RMB (-12% to -59%) |
| Prob-Weighted Return |
-30% to -39% |
All five models independently reached the same conclusion: the stock has front-run fundamentals. Even at the current all-time-high APT price of RMB 810K/ton, forward PE remains 22–45x — far above the historical cycle-peak PE of 13–25x. The risk/reward is asymmetrically skewed to the downside.
1. Company Snapshot
| Item |
Detail |
| Full Name |
China Tungsten and Hightech Materials Co., Ltd. |
| Ticker |
000657.SZ (Shenzhen A-share) |
| Parent |
China Minmetals Corporation (Central SOE) |
| Core Business |
Full tungsten value chain: mining → smelting (APT) → powder → hard alloy → cutting tools |
| Key Asset |
Shizhuyuan mine (world’s largest tungsten mine, consolidated Dec 2024) |
| Capacity |
W concentrate 7,800 t/yr, APT 20,000 t/yr, Hard alloy 14,000 t/yr |
| Resources |
1.23M tons tungsten (11% of China’s reserves) |
| Market Position |
#2 in China tungsten, >70% PCB micro-drill domestic share |
Business Segments (2024 Revenue = 14.743B RMB)
| Segment |
Revenue (B) |
% |
Gross Margin |
APT Sensitivity |
| Concentrate & Powder (APT, WC) |
4.573 |
31% |
23.6% |
Direct (80%+) |
| Other Hard Alloy |
3.356 |
23% |
18.6% |
Lagged 6M, cost squeeze |
| Cutting Tools & Instruments |
3.189 |
22% |
33.3% |
Partial + AI/PCB volume |
| Refractory Metals (W, Mo, Ta) |
2.426 |
16% |
10.3% |
Moderate (40%) |
2. Two-Cycle Backtrack (CRule 1)
2.1 Cycle Identification
| Feature |
Cycle 1 (2010–2015) |
Cycle 2 (2016–2023) |
Cycle 3 (2024–Present) |
| Driver |
Global commodity super-cycle (demand) |
Environmental crackdown + supply reform |
Export controls + AI + military (supply shock) |
| APT Peak |
~200K RMB/ton (2011) |
~170K RMB/ton (2018) |
810K RMB/ton (Feb 2026) — ALL-TIME HIGH |
| APT Trough |
~80K (2015) |
~140K (2020–23) |
N/A (still rising) |
| Stock Peak |
~35 RMB |
~15 RMB |
66.2 RMB (current) |
| Peak PE |
13–25x |
20–25x |
135x |
| Peak PB |
2–3x |
2.2–3x |
15x |
| Peak-to-Trough Stock |
-86% |
-47% |
??? |
2.2 Five-Phase Cycle Anatomy
| Phase |
Rate |
Stock |
PE |
Duration |
| ① Trough |
Below breakeven |
Depressed, high/negative PE |
N/A or >40x |
1–2 years |
| ② Early upturn |
Rising past breakeven |
Stock leads rate by 1–3 months |
PE declining |
6–12 months |
| ③ Mid-cycle |
Sustained above average |
Accelerating, sell-side upgrading |
Compressing |
6–18 months |
| ④ Peak/Mania |
Historic highs |
May plateau or lag rate |
PE at cycle lows |
3–6 months |
| ⑤ Downturn |
Declining |
Stock leads decline by 1–3 months |
Expanding |
1–3 years |
2.3 Current Position — Model Consensus
| Model |
Current Phase |
Evidence |
Time to Peak |
| Opus 4.6 |
Phase ④ Euphoria |
Stock leverage 2x APT; 135x PE » historical |
At or near peak |
| Sonnet 4.6 |
Late Phase 4, early Phase 5 |
600% gain; PB 15x vs 2–3x history |
Weeks to months |
| GPT-5.2 |
Phase 4 (Mania/Crowded) |
EPS lag; consensus has not caught up |
3–6 months |
| GPT-5.1 |
Phase 4 (Mania) |
135x PE unsustainable; awaiting PE compression |
0–3 months |
| Gemini 3 Pro |
Phase 4 (Mania/Squeeze) |
Stock coincident with spot; any hesitation = crash |
At peak |
★ Consensus: Deep in Phase 4 (Mania/Euphoria). Stock has front-run APT by 2x leverage. Historical pattern: stock peaks 1–2 quarters after APT peak, then declines 1.5–2x the APT decline.
3. PE Compression Path
Earnings Catch-Up (Stock Price Unchanged at 66.2)
| Required PE |
Required NI (B RMB) |
Required EPS |
vs 2024 NI |
| 50x |
3.02B |
1.33 |
+222% |
| 30x |
5.03B |
2.21 |
+436% |
| 25x |
6.04B |
2.65 |
+543% |
| 20x |
7.55B |
3.31 |
+703% |
| 15x |
10.07B |
4.42 |
+971% |
★ Even at super-bull earnings (NI 5–10B), forward PE remains 15–30x — still above most models’ assigned peak PE.
4. Operating Leverage (CRule 4)
| APT (K RMB/ton) |
vs Breakeven |
Mine Margin |
Profit Multiple vs 2024 |
Gross Margin |
| 200 (2024 avg) |
1.0x |
~100K |
1.0x |
~50% |
| 400 (Bear) |
2.0x |
~300K |
3.0x |
~75% |
| 600 (Consensus) |
3.0x |
~500K |
5.0x |
~83% |
| 810 (Spot) |
4.05x |
~710K |
7.1x |
~88% |
| 1,000 (Bull) |
5.0x |
~900K |
9.0x |
~90% |
| 1,200 (Super-bull) |
6.0x |
~1,100K |
11.0x |
~92% |
5. 2026E Earnings Sensitivity (CRule 7) — Multi-Model
5.1 Model Results
| APT (K/ton) |
Opus 4.6 |
Sonnet 4.6 |
GPT-5.2 |
GPT-5.1 |
Gemini 3 |
Average |
| 400 (Bear) |
2.18B |
~2.0B |
2.65B |
1.30B |
2.17B |
2.06B |
| 600 (Cons) |
2.90B |
~3.0B |
4.58B |
1.66B |
3.34B |
3.10B |
| 810 (Spot) |
3.67B |
~4.2B |
6.61B |
2.04B |
4.77B |
4.26B |
| 1,000 (Bull) |
4.35B |
~5.2B |
8.19B |
2.38B |
5.88B |
5.20B |
| 1,200 (Super) |
5.08B |
~6.5B |
9.93B |
2.74B |
7.15B |
6.28B |
5.2 Forward PE at Current Price (66.2 RMB)
| Scenario |
Avg NI (B) |
Avg EPS |
Forward PE |
vs Historical Peak (13–25x) |
| Bear (400K) |
2.06B |
0.90 |
73x |
3–6x above |
| Consensus (600K) |
3.10B |
1.36 |
49x |
2–4x above |
| Spot (810K) |
4.26B |
1.87 |
35x |
1.4–2.7x above |
| Bull (1,000K) |
5.20B |
2.28 |
29x |
~At top of range |
| Super-bull (1,200K) |
6.28B |
2.76 |
24x |
Within range |
6. Target Prices — Multi-Model
12-Month Targets
| Scenario |
Opus 4.6 |
Sonnet 4.6 |
GPT-5.2 |
GPT-5.1 |
Gemini 3 |
Consensus |
| Bear |
17 |
~18 |
21 |
11 |
29 |
19 (-71%) |
| Base |
45 |
55–72 |
58 |
18 |
52 |
46 (-30%) |
| Bull |
67 |
~90 |
90 |
27 |
94 |
74 (+12%) |
Probability-Weighted Expected 12M Return
| Model |
Expected Return |
| Opus 4.6 |
-35% |
| Sonnet 4.6 |
-20% to -40% |
| GPT-5.2 |
-30% to -40% |
| GPT-5.1 |
-60% to -80% |
| Gemini 3 Pro |
-21% |
| Consensus |
-30% to -39% |
7. Day1Global Framework
Module Grades
| Module |
Focus |
Grade |
Key Issue |
| C (Cash Flow) |
FCF, capex, working capital |
B |
Mining FCF surge at 810K, but inventory/WC risk |
| L (Ownership) |
SOE structure, governance |
C+ |
China Minmetals; SOE discount; related-party risk |
| O (Accounting) |
Revenue quality, inventory |
C+ |
Inventory revaluation risk; high related-party exposure |
8. Six Investment Perspectives
| Perspective |
Verdict |
One-Liner |
| ① Buffett |
AVOID |
135x PE on cyclical commodity = speculation, not investment |
| ② Baillie Gifford |
Interested but overpriced |
AI/PCB thesis is real growth; wrong vehicle at this price |
| ③ Tiger Cubs |
6 months too late |
+600% done; short side more interesting now |
| ④ Klarman |
SHORT/AVOID |
PB=15x, no margin of safety, every assumption must go right |
| ⑤ Tepper |
Wait for the bottom |
Buy cyclicals at the bottom (5–10 RMB), not the top |
| ⑥ Druckenmiller |
Trimming |
Would have been long at 10; now trailing stop at 50 |
9. Pre-Mortem Analysis
| Scenario |
Probability |
What Happened |
Resulting Price |
| A: APT Collapse |
30–35% |
Export controls leaked; non-China supply; AI normalized. APT → 400K. |
15–20 RMB |
| B: PE Compression |
25–30% |
APT stayed 600–700K (good!), but market de-rated to 20–30x. Classic “beat and sell.” |
30–40 RMB |
| C: Gov Intervention |
10–15% |
Strategic reserves released; price cap ~500K; windfall tax. |
25–30 RMB |
| D: Macro Shock |
15–20% |
Global recession; metals collapse; A-share -30%; high-beta -60%. |
20–26 RMB |
Combined probability of ≥40% loss within 12 months: ~55–65%
10. Risk Matrix
| Risk |
Probability |
Impact |
Severity |
| APT mean-reversion (>40%) |
40% |
Very High |
🔴 |
| PE compression (beat & sell) |
60% |
High |
🔴 |
| Export control reversal |
20% |
High |
🟠 |
| Government price cap |
15% |
High |
🟠 |
| AI/PCB demand miss |
25% |
Medium |
🟡 |
| SOE M&A misallocation |
25% |
Medium |
🟡 |
| Non-China supply (3–5yr) |
30% |
Medium |
🟡 |
| Inventory write-down |
35% |
Medium |
🟡 |
11. Investment Recommendation
Rating: SELL / TAKE PROFIT (5/5 consensus)
| Profile |
Action |
| Currently Long |
Sell 50–75%. Trail stop at 50 RMB on remainder. |
| Considering Entry |
Do not buy. Wait for 35–40 RMB (25–30x fwd PE). |
| Short Sellers |
Attractive R/R; consider put spreads (borrow expensive). |
| Cyclical Traders |
Backwardation = hold; contango = exit. |
What Would Change Our Mind?
- APT sustains >800K for 12+ months → Upgrade to HOLD
- Stock pulls back to 35–40 RMB → Upgrade to BUY
- Non-China W supply permanently impaired → Raise PE ceiling
- Cutting tools GM >40% sustained → Higher base NI
12. Peers
| Company |
Ticker |
2024 Rev (B) |
Notes |
| 厦门钨业 |
600549 |
~35.3 |
Largest; diversified W + rare earth + battery |
| 中钨高新 |
000657 |
14.7 |
Pure tungsten chain; PE 135x |
| 章源钨业 |
002378 |
~3.5 |
Third; resource-type, own mines |
Appendix: Model Rating Summary
| Model |
Rating |
12M Base Target |
Key Insight |
| Opus 4.6 |
SELL |
44 RMB (-33%) |
Prob-weighted -35%. Only super-bull justifies price. |
| Sonnet 4.6 |
HOLD/SELL |
55–72 RMB (-12%) |
Late Phase 4. Stock peaks 2–4 months before APT. |
| GPT-5.2 |
Neutral |
58 RMB (-12%) |
Detailed segment model; PE compression ahead. |
| GPT-5.1 |
SELL/AVOID |
18 RMB (-73%) |
Most bearish. Linear rent model shows massive downside. |
| Gemini 3 Pro |
SELL/TAKE PROFIT |
52 RMB (-21%) |
“Tech multiple for a cyclical miner at peak.” |
Report generated March 4, 2026. Analysis powered by 5 AI models.
This report does not constitute investment advice.