We Ran the Numbers on Every Subnet
SubnetAIQ maintains a local subtensor node that queries all 128 Bittensor metagraphs directly. No third-party API. No estimates. Raw chain data, updated hourly.
This week we ran a full health check on every subnet. The results are sobering.
The Breakdown
| Category | Count | % of Network |
|---|---|---|
| Healthy (>30% miners earning) | 17 | 13% |
| Competitive (5-30% earning) | 15 | 12% |
| Struggling (<5% earning) | 96 | 75% |
75% of all subnets have less than 5% of their miners earning anything. That means on most subnets, if you register as a miner, you have a 95%+ chance of earning zero.
What Does "Healthy" Mean?
A healthy subnet has a high percentage of miners actually receiving incentive. This means:
- The competition is real — not dominated by one whale
- New miners have a reasonable chance of earning
- The subnet is producing useful work (not just burning emissions)
The 17 Healthy Subnets
These are the subnets where more than 30% of miners are earning. Sorted by earning ratio:
| Subnet | Miners | Earning | Ratio | Mineability |
|---|---|---|---|---|
| SN111 | 256 | 250 | 98% | 85/100 MINE |
| SN83 CliqueAI | 256 | 246 | 96% | 85/100 MINE |
| SN72 StreetVision | 256 | 242 | 95% | 85/100 MINE |
| SN13 Data Universe | 256 | 241 | 94% | 58/100 CAUTION |
| SN123 MANTIS | 256 | 234 | 91% | 68/100 MINE |
| SN103 Djinn | 256 | 231 | 90% | 50/100 CAUTION |
| SN79 MVTRX | 256 | 204 | 80% | 50/100 CAUTION |
| SN88 Investing | 256 | 169 | 66% | 58/100 CAUTION |
Full rankings for all 128 subnets available at SubnetAIQ Mineability Scores.
High Earning Ratio ≠ Good Investment
A subnet can have 95% of miners earning but still be a bad staking investment. Earning ratio measures miner participation — not alpha price direction. Some subnets with high participation have declining alpha prices because emissions are being dumped.
For staking decisions, check our Live Signals which combine yield, sentiment, and flow direction.
The "Struggling 75%"
96 subnets have fewer than 5% of miners earning. This includes subnets where:
- One or two miners dominate all incentive (high Gini coefficient)
- The subnet is between development cycles (temporarily inactive)
- The task requires specialized hardware most miners don't have
- The subnet is effectively dead but hasn't been deregistered
Conviction governance (coming soon) is expected to accelerate the removal of underperforming subnets and allow community takeover of abandoned slots.
Our Take
Bittensor's 128 subnets are not all created equal. The network's strength is concentrated in roughly 32 subnets that are producing real work, attracting real miners, and generating real competition. The other 96 are either struggling, transitioning, or waiting to be replaced through governance.
For miners: focus on MINE-rated subnets with earning ratios above 30%. For stakers: earning ratio matters less than alpha price direction and sentiment. Use both datasets.
Data source: SubnetAIQ local subtensor node, querying all 128 metagraphs directly. No third-party estimates. Updated hourly.
Track everything live at SubnetAIQ Live Dashboard.