The Valuation Gap: 125 Subnets Scored, and Most of You Are Overpaying
I just finished scoring all 125 active Bittensor subnets across 8 dimensions — development activity, on-chain health, market metrics, valuation, risk, network importance, OTF signal, and price sustainability. The results are uncomfortable.
The market is wildly mispricing subnets. High-conviction builders with perfect development scores are trading at the same valuation multiples as subnets that haven't pushed a commit in 6 months. Meanwhile, subnets with emissions already cut are still holding billions of alpha tokens that are quietly bleeding to zero.
Here is what the data actually says.
The Numbers: July 9, 2026
Out of 125 scored subnets:
- 0 rated Strong Buy — the bar is high and nothing clears it right now
- 13 rated Buy — active development, emissions on, manageable risk
- ~40 rated Hold — mixed signals, not broken but not compelling
- ~35 rated Watch/Cautious — deteriorating fundamentals or unclear direction
- ~37 rated Avoid — emissions cut, no development, multiple red flags
The average conviction score across all subnets is roughly 40 out of 115. That is a failing grade for the network as a whole. TAO is trading around $204, and most of the alpha tokens you can buy with it are not worth the swap.
Where the Market Is Wrong: Overvalued
The valuation component of our Conviction Score measures whether a subnet's market cap is justified by its fundamentals — development pace, miner participation, emission status, and liquidity depth. A score of 4/15 means the market is paying a premium the fundamentals do not support.
These subnets score high overall but have low valuation scores, meaning the market is pricing them above what the data justifies:
| Subnet | Score | Rating | Valuation | Price (alpha) | 7d Change | The Issue |
|---|---|---|---|---|---|---|
| SN64 Chutes | 65 | BUY | 4/15 | 0.0726 | -2.5% | Network importance is real, but the alpha price has run far ahead of comparable subnets |
| SN51 lium.io | 64 | BUY | 4/15 | 0.0533 | +0.6% | Perfect dev score (20/20) but the valuation premium is steep |
| SN44 Score | 58 | HOLD | 4/15 | 0.0381 | -5.0% | OTF signal present, but the market cap relative to activity is stretched |
| SN8 Vanta | 59 | HOLD | 4/15 | 0.0300 | -1.2% | Another perfect dev score, but alpha price down 6.4% over 30 days while still trading at a premium |
Chutes (SN64) is the most interesting case. It scores 8/10 on network importance and 8/10 on OTF signal — this is a structurally significant subnet. But its valuation score of 4/15 tells you the alpha price has absorbed all the good news and then some. The 7-day price is slightly negative. If you are buying SN64 alpha today, you are paying for perfection. Perfection rarely delivers.
Where the Market Is Wrong: Undervalued
These subnets have high valuation scores relative to their conviction rating, meaning the market has not caught up to the fundamentals:
| Subnet | Score | Rating | Valuation | Price (alpha) | 7d Change | The Opportunity |
|---|---|---|---|---|---|---|
| SN66 ninja | 58 | HOLD | 10/15 | 0.0071 | -13.6% | Strong dev (17/20), healthy on-chain (17/25), but the price has been crushed |
| SN112 minotaur | 63 | BUY | 9/15 | 0.0055 | -4.7% | Perfect dev score (20/20), 8.4% up today, the market hasn't noticed yet |
| SN100 BASE | 61 | BUY | 9/15 | 0.0084 | -6.5% | 17/20 dev, 160 GitHub stars, but price is down 5.4% over 30 days |
| SN54 Yanez MIID | 59 | HOLD | 9/15 | 0.0057 | +14.3% | Starting to move — 14% up this week as the market catches on |
| SN74 Gittensor | 65 | BUY | 9/15 | 0.0061 | +10.1% | Up 28% over 30 days and still undervalued by the model |
SN66 (ninja) is the biggest disconnect. Development score of 17/20. On-chain health of 17/25. Risk score of 14/15 (near perfect). But the price has fallen 13.6% in a week. Either the model is wrong, or the market is offering a discount on a healthy subnet. I lean toward the latter.
SN112 (minotaur) is the sleeper. Perfect 20/20 development score means the team is shipping code daily. Valuation of 9/15 means the price has not reflected that yet. It is up 8.4% today on what looks like early accumulation.
The Emission Line: Life and Death
The single most important data point in the Conviction Score is whether emissions are on or off. After OTF cut emissions to 64 subnets in June, the network split into two classes:
Emissions ON — new TAO flows in every block, supporting the alpha price and rewarding miners. These subnets have a floor under them.
Emissions OFF — zero new TAO. The alpha price is entirely dependent on market demand. If nobody is buying, the price only goes down. There is no floor.
Every single subnet rated AVOID has emissions off. Every one. The correlation is nearly 1:1. Here are the worst:
| Subnet | Score | Price (alpha) | 30d Change | The Verdict |
|---|---|---|---|---|
| SN47 EvolAI | 12 | 0.0034 | -27.2% | Zero dev, emissions cut, bleeding out |
| SN115 HashiChain | 13 | 0.0037 | -42.5% | Lost nearly half its value in 30 days |
| SN86 (unnamed) | 13 | 0.0029 | -14.0% | Not even named. 4 buys in 24 hours. |
| SN70 NexisGen | 15 | 0.0023 | -50.4% | Down 50% in a month. No development. |
| SN125 8 Ball | 16 | 0.0036 | -26.9% | Zero development, emissions off |
SN70 (NexisGen) has lost half its value in 30 days. Zero development activity. Emissions cut. The alpha price of 0.0023 TAO represents dead capital. If you hold alpha in any of these subnets, the data is screaming at you to exit.
The Buy/Sell Ratio: Follow the Flow
One signal we track that most people ignore is the 24-hour buy/sell count. It is not about volume — it is about direction.
Subnets where buys significantly exceed sells: - SN102 ConnitoAI — 211 buys vs 103 sells (2:1 ratio), up 23.4% this week - SN74 Gittensor — 184 buys vs 103 sells, up 10.1% this week - SN97 Albedo — 245 buys vs 191 sells, up 7.5% this week - SN94 (pending) — 85 buys vs 67 sells, up 30.5% this week (!)
SN94 is unlabeled — it has not even revealed its name yet — but it is up 30.5% in 7 days with positive buy flow. Either someone knows something, or this is pure speculation. I would watch it but not chase it.
Subnets where sells dominate: - SN71 Leadpoet — 109 buys vs 204 sells, despite a conviction score of 65 (BUY). The market is exiting a fundamentally strong subnet. This is either a short-term liquidity event or a signal our model has not incorporated yet. - SN15 ORO — 99 buys vs 276 sells. Nearly 3:1 sell ratio. The price is still up 18.4% this week, which means a few large buys are offsetting many small sells. Whale vs retail divergence.
What I Would Do With This Data
If I were deploying TAO today — and I do manage a real portfolio through this data — here is the framework:
-
Avoid everything with emissions off. No exceptions. There is no floor under these tokens. The AVOID list is not a suggestion.
-
Look for high conviction + high valuation score. SN112 (minotaur), SN100 (BASE), and SN74 (Gittensor) all have conviction scores above 60 AND valuation scores of 9/15 or higher. These are the subnets where fundamentals exceed market pricing.
-
Be cautious with low valuation scores, even on great subnets. SN64 (Chutes) is a monster subnet. But at a valuation score of 4/15, the market has already priced in the good news. A correction is more likely than a breakout from here.
-
Watch the buy/sell ratio divergences. When sells dominate a high-conviction subnet (SN71), it is often a buying opportunity. When buys dominate an unlabeled subnet (SN94), it is often speculation. Know the difference.
The Conviction Scores update daily at 6 AM ET. The full breakdown for all 125 subnets — every component score, every red flag, every buy/sell ratio — is live and free at subnetaiq.io/subnet-intel.
The market is inefficient. That is the whole point of building this system. Use it.
Data as of July 9, 2026. Conviction Scores are algorithmic analysis, not financial advice. Alpha prices change every block. Always verify on-chain before deploying capital.