The ultimate question:"In this system, who is the silent, irreplaceable physical switch?"
Trigger method
- User says: "Analyze XXX using the Serenity method"
- User says: "Help me find the chokepoints in the XXX industry chain"
- User says: "Do a supply chain reverse breakdown"
- User mentions "chokepoint / chokepoint / bottleneck / supply chain breakdown"
Three first principles (recite before analysis)
- Don't chase hot topics; walk the narrow gate — Excess returns hide in overlooked small-cap key nodes, not in already fully priced leaders
- Driven by physical iron laws, not narratives — Derive the inevitable path from the physical limits of power consumption, latency, heat dissipation, and bandwidth density; don't listen to stories
- Engineer-style DD, not analyst-style repetition — Dig through primary materials (technical documents/SEC filings/patents/customer roadmaps/job postings), use AI for red team-blue team exercises
Six-step practical workflow
Step 1: Confirm the super-certain megatrend (Demand Validation)
Notchasing short-term hot topics,butjudging whether the underlying demand has reached a "physical-level certainty".
Validation methods:
- Capex guidance in hyperscaler earnings reports (Microsoft/Meta/Google/Amazon/Oracle/tech company cloud, etc.)
- NVIDIA/AMD product roadmaps
- Relevant physical limit papers (power wall, bandwidth wall, thermal wall)
- Policy-mandated phase-out roadmaps (e.g., EU energy efficiency standards)
Key judgment criteria(stop if any is not met):
Does this trend have an irreversible physical or engineering driving force? If it is only "there may be demand", the entire process should stop. Only when you confirm that "downstream must do this, otherwise it is physically unworkable" is it worth digging deeper.
Output: a sentence describing the trend with "physical-level certainty".
Step 2: Reverse-map the complete supply chain (Reverse Mapping)
The most valuable methodological innovation. Break down layer by layer from downstream end products to upstream, rather than looking from upstream to downstream at "whose products sell well".
Practical tools
- Supplier lists in public financial reports
- Industry research reports (Gartner/IDC/Omdia)
- Patent databases (Google Patents, USPTO, China National Intellectual Property Administration)
- Company website customer pages / investor relations pages
- Job postings (infer technology direction from hiring positions)
- Industry exhibition exhibitor lists
- Academic paper acknowledgments section (often lists key equipment/material suppliers)
Reverse breakdown example (taking CPO/optical interconnect as an example)
GPU 集群 / 数据中心 ASC（Trainium / Maia）
→ 网络交换芯片
→ 光模块
→ 外部光源（ELS）/ CW 激光器
→ 衬底材料（InP / SOI）
→ 晶体生长 / 原料（镓、铟、砷）
→ 测试 / 封装设备
Deliverable
A multi-layer supply chain map (can use tables/tree structures/Miro canvas), annotated with:
- Number of suppliers at each link
- Production capacity scale (if public data is available)
- Height of technical barriers
- Capacity expansion cycle (18 months? 3 years?)
- Customer validation status
Step 3: Identify the True Chokepoint (True Chokepoint Checklist)
After locating potential chokepoints on the supply chain map, use the following 5-dimension Checklist for strict screening. All 5 criteria must be satisfied simultaneously for it to be a high-quality chokepoint.
| Dimension | Judgment criteria | Core question |
|---|---|---|
| Constrained Supply | Slow capacity expansion, difficult to expand capacity | Long qualification cycle? High yield barrier? Physical/chemical limits? Is the capacity expansion cycle 18 months or 3 years? |
| Irreplaceable | Very few global suppliers (1-3) | How high is the cost for downstream to switch suppliers? Are there alternative solutions along the technology route? |
| Low Attention | Small market cap, little institutional coverage, poor liquidity | How many analysts on Wall Street/brokerages cover it? How large is the average daily trading value? How heated is market discussion? |
| High Value Capture | Pricing power, high profit elasticity | Can the supply-demand mismatch translate into profit growth? What is the gross margin trend? |
| Catalyst Visible | Imminent mass production, qualification passed, financial report verification | What upcoming events can verify the investment thesis? M&A? Index inclusion? |
Beware of "fake chokepoints"
Companies in the supply chain that "also have a share" but are replaceable, lack scarcity, and lack pricing powerare nottrue chokepoints.
The ability to distinguish true from fake chokepoints is the most core judgment in this methodology.
Step 4: Layered Evidence Verification + Anti-Thesis Stress Test (Evidence & Anti-Thesis)
Evidence layering system (decreasing weight)
| Tier | Signal strength | Specific type | Usage rules |
|---|---|---|---|
| High-tier evidence (strong signal) | ★★★★★ | Direct confirmation by management in company announcements, SEC filings, earnings calls, customer contracts or roadmaps | Can directly verify the supply chain logic and be used alone as a basis for decision-making |
| Mid-tier evidence (medium signal) | ★★★☆☆ | Patent data, recruitment information (inferring technology direction from job postings), supplier lists, government-funded projects | Can indirectly support the logic, but cannot be used alone as a basis for decision-making |
| Low-tier evidence (weak signal) | ★☆☆☆☆ | Peer mapping, speculative analysis, social media discussion | Can only serve as auxiliary reference and must be used with caution |
Anti-Thesis Table — the most valuable part
For each investment thesis, proactively build an "anti-thesis table".
Standard anti-thesis question template(copy into your analysis document and fill it out):
【反论表：针对 ______ 的批判性审查】
1. 如果客户不放量怎么办？
- 量化：如果需求只有预期的 ___%，利润影响是多少？
- 对策：______
2. 如果竞争者绕过这个技术路线怎么办？
- 替代技术：______
- 替代时间窗口：______
- 对策：______
3. 如果估值已经提前反映了预期怎么办？
- 当前市值已反映 ___ 年的增长？
- 安全边际：______
4. 如果出现稀释或财报 miss 怎么办？
- 最大稀释比例：______
- 历史 miss 频率：______
5. 如何证伪这个论点？（Falsifiability）
- 可观测的证伪信号：______
- 如果 ______ 发生，说明这个论点错了
AI red-team/blue-team confrontation (practical steps)
- Organize your complete investment logic (the output of Steps 1-3) into a piece of text
- Send it to AI (Gemini/Claude/WorkBuddy) along with the following Prompt:
你现在是我的"红队"（Red Team）。请对我的以下投资逻辑进行最严厉的批判性审查。
我的论点是：
[粘贴你的完整逻辑]
请从以下角度攻击我的论点：
1. 技术替代路径（有没有办法绕过这个瓶颈？）
2. 产能核实（声称的产能是否真实？）
3. 客户验证（是否有独立的客户确认？）
4. 估值透支（当前价格反映了多少年的增长？）
5. 地缘政治风险（供应链是否依赖单一国家？）
6. 管理层诚信（是否有历史问题？）
7. 流动性风险（日均成交量是否足够退出？）
对每个角度，请给出具体的质疑点，并建议我如何验证或反驳。
- Respond to each of AI's challenges one by one; for those that cannot be answered, mark them as "risks pending verification"
- Repeat for 2-3 rounds until no new valid challenges arise
This "attacking yourself" mode of thinking is the key dividing line between amateur analysis and professional investment research.
Step 5: Catalyst and Timing Judgment (Catalyst Timing)
It is notholding indefinitely once you find a bottleneck,but ratherwaiting for short-term catalysts to materialize.
Catalyst Type Checklist
| Type | Specific Event | Verification Method |
|---|---|---|
| Order Announcement | Major customer procurement announcement, long-term supply agreement | Company announcements, SEC filings |
| Product Certification | Industry certification passed, customer validation completed | Company press releases, customer official website |
| Financial Report Verification | Revenue/profit inflection point appears | Quarterly financial reports, earnings call |
| Big Tech Investment/Partnership | Hyperscaler strategic investment, joint R&D | News, SEC filings |
| M&A Event | Acquired by a large company | Company announcements |
| Index Inclusion | Included in major indices (S&P/Nasdaq/CSI 300) | Index company announcements |
Institutional Rotation Signals (External Confirmation)
When major institutions begin building positions or publishing related research, it often means market perception is catching up; this is external confirmation that the logic is correct.
Tracking methods:
- 13F filings (quarterly position disclosures by US institutions)
- Sudden coverage by brokerage research reports (especially initiation of coverage)
- Block trade records
Timing principle: position before the catalyst becomes clear, and consider reducing positions when the catalyst materializes and market consensus forms.
Step 6: Position Sizing and Risk Management (Position Sizing)
Core principle:Position size should strictly match research depth, not "level of confidence"
| Research Depth | Maximum Allowable Position | Description |
|---|---|---|
| Only read others' summaries and analyses | ≤ 2% | Probe position, to verify your own understanding |
| Drew your own supply chain map + read public materials | ≤ 5% | Initial position |
| Completed Steps 1-4 in full (including counter-thesis table) | ≤ 10% | Core position |
| All of Steps 1-5 completed + multi-source evidence cross-validated | ≤ 15-20% | Full position (a single security must not exceed this limit) |
Stop-loss logic:Based on thesis invalidation, not price fluctuations
| Situation | Action |
|---|---|
| Industry chain logic is invalidated (a change in technology route makes this segment no longer critical) | Exit immediately, even if the price is still rising |
| Logic has not been invalidated, only market sentiment fluctuation | Hold or even add to position |
| A better bottleneck appears (a superior target in the same industry chain) | Rebalance |
| Macro risk event (geopolitical conflict/Sharp change in interest rates) | Temporarily deleverage, but not necessarily liquidate everything |
Leverage Rules
- Normal: ≤ 1.5x leverage
- High-certainty targets validated by multiple sources: ≤ 2x leverage
- During macro risk events: deleverage to ≤ 1x or fully unleveraged
Complete Analysis Output Template
After completing the six-step process, output a structured analysis report containing the following sections:
# [公司名] 供应链卡点分析报告
## 一、超级趋势验证（Step 1）
- 趋势描述：
- 物理级确定性证据：
- 风险：
## 二、供应链地图（Step 2）
[粘贴你的供应链地图]
## 三、真卡点评估（Step 3）
[5维 Checklist 打分表]
## 四、证据与反论（Step 4）
- 证据分层汇总：
- 反论表：
- AI红蓝对抗结论：
## 五、催化剂时间表（Step 5）
[未来6-12个月的关键事件节点]
## 六、投资决策（Step 6）
- 建议仓位：
- 入场区间：
- 止损条件（论点证伪信号）：
- 目标仓位：
## 七、风险提示
[列出所有已识别的风险，按严重程度排序]
7-Day Practice Template (Actionable)
The essence of this methodology is a skill that can be practiced repeatedly, not a one-time "oracle."
| Day | Task | Output |
|---|---|---|
| Day 1-2 | Choose a certain trend and verify its "physics-level certainty" | One-page trend description + 3 pieces of verification evidence |
| Day 3-4 | Draw a supply chain map (decompose from the end product upstream) | Complete multi-layer supply chain map |
| Day 5 | List 5-10 potential chokepoints and score each one using the True Chokepoint Checklist | Checklist scoring sheet |
| Day 6 | Collect and layer evidence, build a counter-argument table, and use AI for red team/blue team adversarial exercises | Evidence layering table + counter-argument table |
| Day 7 | Simulated valuation (current market cap vs. reasonable market cap under a 2-3 year ramp scenario) | Valuation range + margin of safety analysis |
Core insight: Copying code is easy; copying deep insight is hard. The real value of this Skill is not "buying along" but repeatedly practicing the ability to "decompose the chain" and "counter-argue."
Munger Perspective · Critical Usage Guide
When using this Skill, always apply the following Munger mental models for secondary review:
| Mental Model | Review Question |
|---|---|
| Inversion | "If this thesis is wrong, what would be the reason?" |
| Circle of Competence | "Do I really understand this technical link, or do I need to trust blindly?" |
| Margin of Safety | "Does the current price leave a sufficient margin of safety?" |
| Latticework of Mental Models | "Besides supply chain logic, what perspectives from other disciplines can validate or refute this thesis?" (physics/game theory/psychology) |
| Psychology of Human Misjudgment | "Am I overconfident because of recent successful cases? (confirmation bias)" |
Risks Munger Would Particularly Emphasize：
- Liquidity trap — The "price elasticity" of micro-cap stocks is a double-edged sword: they are extremely elastic on the way up, but there may also be no buyers on the way down
- Single technology path dependence — Betting everything on "CPO is the only physical evolution path" violates Munger's "always invert"
- Survivorship bias — You see Serenity's AXTI that won 1000%, but you don't see the positions that could have been wiped out
- "Pump and dump" ethical risk — A person building a position in advance on an anonymous platform, publicly calling the trade, and waiting for retail investors to take over is especially dangerous with micro-cap stocks.If you use the results of this Skill's analysis, never publicly recommend them to others without full disclosure.
Applicable Market Notes
| Market | Notes |
|---|---|
| U.S. stocks | Best suited; micro-cap ecosystem is complete, SEC disclosure requirements are high |
| A-shares | Requires adjustment: focus on mid- and large-cap stocks (avoid liquidity traps), use annual reports/prospectuses instead of SEC filings, and pay attention to the policy tailwinds for "specialized, refined, distinctive and innovative" (SRDI) enterprises |
| Hong Kong stocks | Between U.S. stocks and A-shares; note the impact of short-selling mechanisms on micro-cap stocks |
| Cryptocurrency | Extremely high risk; supply chain logic often does not apply (no physical supply chain), use with caution |
Tool recommendations
| Purpose | Recommended tools |
|---|---|
| Supply chain mapping | Miro, Whimsical, Excalidraw, or plain-text tree structures |
| Evidence collection | SEC EDGAR (U.S. stocks), CNINFO (A-shares), prospectuses, patent search |
| AI red team-blue team exercise | WorkBuddy (this tool), Claude, Gemini |
| Valuation modeling | Excel、Google Sheets |
| Catalyst tracking | Company IR calendars, Seeking Alpha, CNINFO |
Disclaimer
This Skill is only a methodological learning framework and does not constitute any investment advice. The supply chain chokepoint investment strategy is a high-risk strategy (small market capitalization, high volatility, concentrated positions) and is only suitable for investors with a technical background and high risk tolerance. Past performance does not represent future returns. Verify independently and exercise strict risk control.