Here is the translation of the provided text into American English:

---

Five-Step AI Industry Research Framework (Highly Recommended to Save)

Step 1: Verify the Authenticity of Terminal Demand
AI demand cannot be driven solely by emotions; it must be based on specific metrics.
Capital expenditure: Check if customers are truly spending money, not just verbal claims.
Architecture upgrade: Is the existing architecture at a point where an upgrade is necessary?
Customer contracts: Are there solid orders and contracts supporting the demand?
Power supply: Behind computing power consumption, is the electricity support keeping up?

Step 2: Decompose along the industry chain upstream to identify the most rigid supply links.
Once demand is confirmed, see who has the hardest time expanding capacity or replacing existing solutions.
Difficulty in expansion: Who has the highest capacity expansion threshold and longest cycle?
Replacement difficulty: Who’s technology or resources are least easily replaced by other solutions?
Certification cycle: Who has the highest entry barriers and longest customer certification time?
Bottleneck position: The stage that is hardest to expand, hardest to replace, and takes the longest to certify—most likely to become an industry bottleneck.

Step 3: Verify whether the company is truly at the bottleneck position. Not all companies talking about AI are at critical nodes; hard metrics are needed to prove it.
Orders: Are there real orders supporting the company?
Customers: Are the customers core players?
Qualification verification: Has the company passed key certifications?
Partners: Are there collaborations with industry leaders?
Capacity expansion: Can the company’s capacity keep up with bottleneck demands?

Step 4: Judging the market stage is not about stock price movements but about the level of industry validation and expectation diffusion.
Industry validation: Are the technical path and business model already validated?
Institutional participation: Has mainstream capital already entered on a large scale?
Order realization: Have orders shifted from expectations to actual performance?
Expectation diffusion: Have good news spread from a few to the masses?
Still early vs priced in:
"Early" means validation and realization are not complete;
"Priced in" doesn’t mean the company is bad, but that good news has already been fully absorbed by the market.

Step 5: Incorporate capital structure into the trading framework, especially in heavy asset and high-investment segments—financial structure is more important than stories.
Financing costs: How much is the interest on borrowed money, and is the cost pressure high?
Dilution pressure: Will equity financing severely dilute existing shareholders?
Debt structure: The ratio of short-term to long-term debt.
Healthy cash flow: Can the company generate its own cash, or is it always relying on external funding?

Using this research model can basically help you find the most valuable targets in various fields—give it a try if you don’t believe it.