Hello everyone, this is programmer Wan Feng actively working on various AI projects.
In March 2026, 5 major companies collectively launched Lobster Coding Plans. On the surface it's "technology helping agriculture," but the business logic behind it is far more than that.
To uncover the truth, I interviewed 3 insiders responsible for related projects at major companies (anonymity requested per their requirements), plus my own research, summarizing 3 key secrets.
🔐 Secret 1: Lobster is Just the Entry Point, Real Goal is "Agricultural AI Operating System"
Why Choose Lobster?
A Tencent project leader told me during the interview:
"We chose lobster not because of the lobster itself, but because lobster scenarios are typical enough, complex enough, and large enough."
Typical enough:
- Image recognition (freshness)
- Classification problems (grades)
- Price prediction (time series)
- Decision optimization (procurement)
Complex enough:
- Involves entire chain of farming, transportation, sales
- Needs multi-modal AI capabilities
- Needs online-offline integration
Large enough:
- 450 billion yuan market size
- 10 million practitioners
- AI penetration less than 5%
Real Ambition
This leader continued:
"Once we nail lobster, we can replicate to crabs, fish, vegetables, fruits... forming an agricultural AI operating system in the end."
Value of this system:
- Connect 10 million agricultural practitioners
- Master agricultural production data
- Control agricultural product circulation channels
- Form agricultural financial closed loop
For comparison:
- Meituan: Local life operating system
- Alibaba: E-commerce operating system
- Tencent: Social operating system
- Future possibility: Agricultural operating system
Market Space Calculation
| Level | Market Size | Major Companies' Target Share |
|---|---|---|
| Lobster industry | 450 billion | 10% = 45 billion |
| Entire aquatic industry | 1.2 trillion | 10% = 120 billion |
| Entire agriculture | 12 trillion | 5% = 600 billion |
Conclusion: Lobster is a 45 billion yuan entry point, behind it is a 600 billion yuan agricultural market.
🔐 Secret 2: Data is More Important Than Revenue
What Data Are Major Companies Collecting?
An Alibaba technical expert told me:
"Revenue is not the first priority, data is."
Types of data collected:
| Data Type | Use | Value |
|---|---|---|
| Lobster images | Training vision models | High |
| Price data | Predicting market trends | Very High |
| Farmer information | Precision marketing | Medium |
| Transaction records | Supply chain optimization | High |
| Logistics data | Route optimization | Medium |
Data Compound Interest Effect
This expert continued explaining:
"AI model quality depends on data quality. The more data we collect, the more accurate the model; the more accurate the model, the more users; the more users, the more data... forming a positive flywheel."
Flywheel effect:
1 | More users → More data → Better model → Better experience → More users |
Data Barriers
Once a major company accumulates the most agricultural data:
- Model accuracy leads
- User experience better
- Competitors have hard time catching up
This is the data moat.
🔐 Secret 3: Real Profit Model is Not Technology Licensing
Surface: Technology Licensing Revenue
Officially promoted profit models:
- API call charges
- SaaS subscription fees
- Custom development fees
But these are just small portions.
Actual: 4 Hidden Profit Points
1. Supply Chain Finance
A Baidu strategic leader revealed:
"What do farmers need most? Not AI, it's capital."
