Step 13 of 14

Entering a Subnet

You found a subnet that looks promising. Now what? This guide walks you through the complete workflow from initial research to a final recommendation for your team lead. Every step is explained in detail so you always know what to do next.

1. The Complete Workflow

Here is the full pipeline, from start to finish. Every step has its own section below with detailed instructions.

1. Research in Nexus — Get a MINE verdict on a promising subnet.
2. Analyze the code — Deep dive the repo to understand scoring and requirements.
3. Estimate profitability — Will it pay more than it costs?
4. Prepare a recommendation — Write up everything your team lead needs to decide.
5. Your team lead reviews and decides — He handles the actual deployment (for now).
6. Monitor — Check if the miner is earning after deployment.
7. Optimize or exit — Improve performance or move to a better subnet.

Don't worry about memorizing all of this at once. Each step builds on the previous one, and you'll get faster with practice.

2. Step 1: Research with Nexus

Every subnet analysis starts in Nexus Intelligence. This is where you find out which subnets are worth investigating further.

How to open Nexus

1. Open Agent Deck on the VPS.
2. Go to the Nexus section.
3. Start a chat conversation.

What to ask Nexus

You're looking for subnets with a MINE verdict — meaning Nexus has analyzed the data and thinks the subnet is worth entering. Here are example prompts you can use:

What to check in the results

3. Step 2: Analyze the Code

Once Nexus gives you a MINE verdict, you need to understand how the subnet actually works. This means reading the code. Don't panic — Claude Code does the heavy lifting.

How to start the analysis

1. Open Agent Deck → Work section.
2. Create a new session (press N).
3. Name it after the subnet (e.g., "subnet-111").

Clone the repo

In Claude Code, ask it to clone and explain the subnet:

Map the codebase with GSD

Once the repo is cloned, use GSD to systematically analyze it:

Then ask Claude Code these critical questions, one at a time:

What you're looking for

4. Step 3: Estimate Profitability

Before recommending a subnet, you need to know if it will actually make money. Here is the basic formula:

Example calculation

Key things that can change the numbers:

• Your miner's score — Higher score = bigger share of emissions
• TAO price fluctuation — If TAO drops 50%, your revenue drops 50%
• New miners entering — More competition = smaller slice per miner
• Subnet emission changes — Staking shifts can increase or decrease a subnet's total emission

5. Step 4: Prepare a Recommendation

This is where everything comes together. You've done the research, analyzed the code, and estimated profitability. Now you write a clear recommendation for your team lead.

What a recommendation should include

1. Subnet number and description — What it does in plain language.
2. Scoring system summary — How miners are judged and what determines top scores.
3. Hardware requirements — What GPU/CPU is needed, and estimated monthly cost.
4. Estimated monthly TAO earnings — Based on current emission and miner count.
5. Risk assessment — Gini, competition level, community health, code quality.
6. Your verdict — MINE with confidence level, or WATCH with what to monitor.
7. Concerns or unknowns — Anything you couldn't verify or that worries you.

Recommendation template

Use this structure when writing your recommendation. Copy it and fill in the blanks:

SUBNET RECOMMENDATION
=====================

Subnet: #[number] - [name]
Date: [today's date]
Verdict: [MINE / WATCH / SKIP]
Confidence: [HIGH / MEDIUM / LOW]

WHAT IT DOES
[1-2 sentences explaining the subnet's purpose]

SCORING SYSTEM
- Miners are evaluated by: [criteria]
- Top scores go to miners who: [what makes a good miner]
- Competition type: [model quality / data quality / compute / hybrid]

HARDWARE REQUIREMENTS
- Minimum: [GPU/CPU type]
- Recommended: [GPU/CPU type]
- Estimated cost: $[amount]/month

PROFITABILITY ESTIMATE
- Emission per miner: [amount] TAO/day
- Current miner count: [number]
- Estimated monthly earnings: [amount] TAO (~$[amount])
- Monthly cost: $[amount]
- Estimated profit: $[amount]/month

RISK ASSESSMENT
- Gini coefficient: [value] ([fair / concentrated / highly concentrated])
- Competition: [low / medium / high]
- Community: [active / moderate / inactive]
- Code quality: [well-maintained / average / abandoned]

CONCERNS
- [List anything unclear, risky, or that needs monitoring]

RECOMMENDATION
[2-3 sentences summarizing why you recommend MINE, WATCH, or SKIP]

6. What Happens Next

After you submit your recommendation, here's how the process continues:

1. Your team lead reviews your recommendation — He checks your reasoning, asks questions if needed, and makes the final call.
2. If he agrees, he handles deployment — For now, your team lead sets up the hardware, configures the miner, and launches it on the subnet.
3. You monitor the results — After deployment, you check Nexus to see if the miner is earning as expected.
4. You recommend adjustments — If performance is below expectations, you investigate why and suggest changes.

Think about it this way: a doctor diagnoses the problem, and a surgeon operates. Right now, you're the doctor — your diagnosis (the recommendation) is what determines the treatment (which subnet to mine). Both roles are essential, and yours comes first.