The Bittensor Top 10 Index: Betting on the Market, Not One Subnet
A market-cap-weighted index of Bittensor's 10 largest subnets — and why owning the basket is a more honest bet on the subnet market than trying to pick the single subnet that wins.
BitExplorer · Jul 30, 2026
Chutes
Rank #2
τ0.0757
-3.64%
lium.io
Rank #1
τ0.0828
-10.56%
Targon
Rank #3
τ0.0561
+9.70%
Actual
Rank #7
τ0.0461
+1.80%
Score
Rank #4
τ0.0388
-11.19%
Affine
Rank #6
τ0.0534
+8.72%
iota
Rank #8
τ0.0293
-12.43%
engy
Rank #5
τ0.03
+21.34%
Minos
Rank #11
τ0.0594
-14.31%
Vanta
Rank #12
τ0.028
+12.99%Ask ten people which Bittensor subnet is going to matter most in two years and you'll get ten different, confident answers. That's not a knock on any of them — it's just what happens in a market with 128 permissionless, independently competing subnets, where the winners aren't obvious in advance even to people paying close attention. Picking the single subnet that wins is genuinely hard. Owning the subnets the market has already voted are the biggest, in proportion to how much it's voted, is a different and considerably more honest bet.
That's what an index is. This is a live snapshot of what a market-cap-weighted top-10 Bittensor subnet index looks like right now, and the case for why it's a reasonable way to express a bullish view on the subnet market as a whole rather than a view on any one of its parts.
Why index the subnet market at all
Every individual subnet carries risk that has nothing to do with whether Bittensor itself succeeds: it can be outcompeted by a better-executed rival in the same category, it can fail to keep validators and miners engaged, or — in the extreme case — it can be deregistered entirely if it stops earning enough root-network weight to hold its slot. None of that is a bet against Bittensor. It's just the normal risk of picking one company in a sector instead of owning the sector.
An index sidesteps the picking problem. Instead of deciding whether Chutes' compute marketplace or Score's computer-vision network is the better long-term bet, you own both, sized by how much conviction the market has already assigned each of them. If one falters, it's a smaller line in a basket, not the entire position. If one breaks out, it was already sized to matter. You're betting that some subnets in the top tier keep compounding — a much easier call than betting on which specific one.
How this index is built
The methodology is the same one most mainstream stock indices use: take the largest constituents by market value, weight each one by its share of the total.
- Rank all active subnets by market cap (circulating alpha supply × current alpha price, in TAO).
- Take the top 10.
- Weight each one by its share of those 10 subnets' combined market cap.
The index, as of July 27, 2026
| Rank | Subnet | Category | Market cap (TAO) | Weight |
|---|---|---|---|---|
| 1 | Chutes (SN64) | Compute marketplace | ~262,300 | 17.3% |
| 2 | lium.io (SN51) | Compute marketplace | ~219,900 | 14.5% |
| 3 | Targon (SN4) | Compute marketplace | ~195,800 | 12.9% |
| 4 | Actual (SN95) | Inference network | ~138,200 | 9.1% |
| 5 | Score (SN44) | Computer vision | ~138,100 | 9.1% |
| 6 | Affine (SN120) | Reasoning / mining | ~136,100 | 9.0% |
| 7 | iota (SN9) | Model training | ~132,300 | 8.7% |
| 8 | engy (SN53) | Verified inference | ~129,600 | 8.5% |
| 9 | Minos (SN107) | Genomics | ~86,500 | 5.7% |
| 10 | Vanta (SN8) | Trading infrastructure | ~79,100 | 5.2% |
View live subnet data on Markets →
Two things jump out. First, these 10 subnets — out of 128 active ones — together account for roughly 45.8% of the entire network's market cap. An index of the top 10 isn't a narrow slice of the market; it's close to half of it. Second, the top 3 alone (all compute marketplaces) make up 44.7% of the index itself — more on why that matters below.
The honest catch: this index is not actually diversified by category
Cap-weighting has a well-known property: it doesn't care what a subnet does, only how much the market has staked into it. Right now, that means three compute marketplaces — Chutes, lium.io, and Targon — make up nearly half the index by weight. If you're drawn to this index because you want broad exposure to "the subnet market," it's worth knowing you're actually taking a fairly concentrated bet on compute-marketplace demand specifically, with genomics, computer vision, trading infrastructure, and model training as smaller satellite positions around it. Bittensor subnet categories explained is worth reading if you want to understand why compute marketplaces tend to cluster at the top — their addressable demand (AI compute broadly) is larger and more external to Bittensor than most other categories.
This isn't a flaw in the method, just something a cap-weighted index always does — the same critique gets leveled at cap-weighted stock indices that end up concentrated in whichever handful of companies are currently largest.
How to actually replicate this today
There's no single "buy the index" transaction on Bittensor yet — each subnet's alpha token is a separate stake. Replicating this index means staking TAO into each of the 10 subnets above in roughly the weights shown — for example, staking 17.3% of your intended position into Chutes, 14.5% into lium.io, and so on down to 5.2% into Vanta. That's 10 separate staking transactions instead of one, and the weights above will already be slightly stale by the time you read this, since alpha prices move continuously.
Rebalancing is on you, not the index
A real index fund rebalances itself. This one doesn't — it's a snapshot, not a maintained product. Left alone, your position will drift from these weights as prices move, the same way any manually-assembled basket drifts from its starting allocation. Whether that's a problem depends on your own view: letting winners grow their weight over time is exactly how cap-weighting is supposed to work, but if you want to hold a fixed allocation, you'd need to periodically check current market caps and rebalance by hand.
Risks, and why this isn't financial advice
- This is a snapshot, not a live product. Membership and weights shift as the market moves — check current numbers on Markets before acting on this table.
- Cap-weighting concentrates you in whatever is already largest — see the category concentration point above.
- Every subnet in this list carries its own individual risk — competition, deregistration, and execution risk don't disappear just because a subnet made the top 10 today. How to evaluate a Bittensor subnet is worth running each constituent through individually, not just trusting its rank.
- Alpha token prices are volatile, same as any single subnet position.
Nothing here is personalized investment advice — it's an explanation of an indexing approach and what it would currently hold, not a recommendation to buy anything. Do your own research before staking.
Want this as a one-click product instead of 10 manual transactions?
In-app buying isn't live on BitExplorer yet — join the whitelist and we'll notify you the moment it opens.
Related Articles
Frequently asked questions
Is there an actual index token or fund for this?
No — this is a conceptual, market-cap-weighted index you can replicate yourself by staking TAO across the 10 subnets in the weights shown, not a single tradeable product. If BitExplorer or anyone else builds an actual one-click index product, the whitelist is the place to hear about it first.
How are the weights calculated?
Each subnet's weight is its market cap (circulating alpha supply × current alpha price, in TAO) divided by the combined market cap of all 10 subnets in the index. It's the same cap-weighting method used by most mainstream stock market indices.
Why not just equal-weight the 10 subnets instead?
You could — it's a legitimate alternative that reduces concentration risk in the largest names. Cap-weighting is the more common default because it lets the market's own conviction (measured in staked TAO) set the weights, rather than an arbitrary equal split, but it does mean the index is more exposed to whichever subnets are already largest.
Does this index update automatically?
Not on its own — the weights and even the membership (which 10 subnets qualify) will drift as market caps change. Replicating this by hand means periodically checking the current top 10 and rebalancing, the same maintenance any manually-tracked index requires.
Is this financial advice?
No. This is an explanation of an indexing approach and a snapshot of what it would currently hold, not personalized investment advice. Alpha token prices are volatile and subnet rankings change — do your own research before staking anything.