Live data: DEPIN (DEPIN) live telemetry · Bittensor (TAO) live telemetry — every value with source, as_of and a verify URL. Get a free $5 API key · try without signup

The gap this closes

Bittensor has been the largest hole in our first-party coverage for months. The subnets are where the economics actually happen — registration auctions, alpha pools, emission splits — but the chain speaks SCALE-encoded substrate storage, not JSON. Until this week, our catalog listed the subnets and could tell you almost nothing about them.

That decoder now exists. As of 2026-08-24, 129 subnets report 13 SUBNET_* metrics each in our on-chain store, polled directly from the public Finney endpoint (entrypoint-finney.opentensor.ai) — no third-party indexer in the path. Latest snapshot in the 7-day window: 2026-08-24T12:07:01Z.

The metric set per subnet:

• SUBNET_NEURONS / SUBNET_MAX_UIDS / SUBNET_ACTIVE_NEURONS — registered seats, seat capacity, and seats doing work
• SUBNET_ALPHA_IN / SUBNET_ALPHA_OUT — the subnet's alpha token pool sides
• SUBNET_TAO_RESERVES / SUBNET_TAO_IN_EMISSION — TAO backing and per-block TAO inflow
• SUBNET_BURN_COST — the current price of a registration seat
• SUBNET_TEMPO / SUBNET_REGISTERED_AT_BLOCK — epoch cadence and age
• SUBNET_MAX_VALIDATORS / SUBNET_VALIDATOR_PERMITS / SUBNET_VOLUME — validator structure and pool volume

Here is what the first full read of the chain says. All figures below were queried live on 2026-08-24 from our store and are checkable against taostats.io/subnets.

Finding 1: seats are full almost everywhere — but "registered" is not "working"

118 of 129 subnets are at their registration cap (SUBNET_NEURONS >= SUBNET_MAX_UIDS). If you want a UID on most subnets, you are not joining an open network; you are paying the burn cost to displace someone. The emptiest subnet, SN36, has 2 of 256 UIDs filled.

The active-neuron column tells a much less crowded story. The busiest subnet by working seats is SN61 with 58 active neurons, followed by SN15 (57) and SN45 (54). At the other end, seven subnets have two or fewer active neurons despite full or near-full registration: SN36 and SN84 at 1; SN103, SN104, SN16, SN26 and SN85 at 2.

That spread — 256 registered, 2 active — is the single most useful diligence signal in the dataset. Registration is a speculative position (a seat you bought); an active neuron is a machine answering queries this tempo. A subnet where the two diverge by 100x is a waiting room, not a workforce.

Finding 2: TAO emission is extremely concentrated

SUBNET_TAO_IN_EMISSION is the per-block TAO flowing into each subnet's pool. Summed across all 129 subnets, our latest reading is 0.1704 TAO/block tracked. The distribution within that total:

• SN107: 0.02365 TAO/block — 13.9% of tracked emission on its own
• SN15: 0.01328
• SN51: 0.01170
• SN64: 0.01063
• ...down a long tail to subnets at effectively 0

Four subnets take roughly a third of the tracked flow. Emission share is the market's live vote on subnet value (it follows alpha pool pricing), so this concentration is the chain telling you where the money believes the useful work is — and how little it believes in the tail.

Finding 3: UID space is not uniform

The standard subnet runs 256 max UIDs, but it is not universal: SN18 runs the largest space at 257, while SN0, SN84, SN68, SN82 and SN104 run reduced 64-UID spaces and SN107 — the emission leader — runs 100. Any per-seat normalization (emission per UID, burn cost per available seat) breaks if you assume 256 everywhere. We store the denominator per subnet so you don't have to.

Method and caveats

Honest limits of this data:

• Source is a single public RPC endpoint. We decode substrate storage from Finney directly. If the OpenTensor entrypoint serves stale state, we inherit it; the as_of timestamp on every row is when we read the chain, not a finality proof.
• "Active" is the chain's definition, derived from subnet consensus state — it means the neuron is participating per protocol accounting, not that its output is any good. Quality-of-work is out of scope for this layer.
• Emission percentages use our tracked total (0.1704 TAO/block) as the denominator — the sum over the 129 subnets we decode, not a protocol constant. Root-network mechanics and rounding mean this can differ slightly from other dashboards' totals; the ranking and shares are the robust part.
• Snapshot cadence is daily-scale polling, not per-block. Burn cost in particular moves intra-day during registration races; treat our value as the level, not the tick.

Verify it yourself

• Independent cross-check: https://taostats.io/subnets — compare neuron counts, max UIDs, and emission share per subnet against the numbers above.
• The raw chain: the same wss://entrypoint-finney.opentensor.ai endpoint we poll is public; SubnetworkN, MaxAllowedUids, and the emission storage maps are readable by any substrate client.
• Our API (every response carries source, as_of, and a verify_yourself URL):

# All 13 subnet metrics for one subnet
curl https://kairossignal.com/v1/network/SN61 -H "X-API-Key: YOUR_KEY"
Rank subnets by a metric
curl "https://kairossignal.com/v1/compare?metric=SUBNET_ACTIVE_NEURONS" \
  -H "X-API-Key: YOUR_KEY"

Full endpoint reference at /api-docs. If you want to poke at it without a key first, /try works in the browser.

For the broader estate this slots into — what first-party telemetry means and which networks have it — see DePIN API access: infrastructure telemetry and the cross-category comparison in WeatherXM to Akash.

---

Data access: free tier at /try, or the $199 Design Partner plan at /pricing for full API access to all subnet metrics with provenance on every row.

---

Try it yourself

Query the live catalog, supply telemetry, and provenance receipts directly: /v1/networks, /v1/supply on the REST API.

Related reading: Bittensor subnet data · DePIN RPC endpoints across chains · DePIN Intelligence guide · DePIN network data: 327 networks, 69 sources

17 of 20 design-partner seats remain at a lifetime-locked $199/mo (full API access, all 19 endpoints, MCP server, Bitcoin-anchored provenance). After seat 20 the price becomes $249/mo. Claim a design-partner seat → · See pricing

---

Get Started With DePIN Intelligence

Kairos Signal provides verifiable, provenance-first telemetry for 327 DePIN networks — 296 with first-party supply data read directly from each network's own API or blockchain. Every value carries a verify_url you can check yourself, and each daily batch is Merkle-rooted and anchored to Bitcoin.

Three ways to access:

Try free — browse networks, supply data, and provenance with no signup. See exactly what you get before paying a cent.

Design Partner — $199/mo forever — full API access, all 19 endpoints, all intelligence engines. Price locked FOREVER for the first 20 partners. 3 of 20 spots claimed, 17 remaining. After 20 fill: $249/mo. Lock your rate →

Pay-per-query via MCP — autonomous agent access. Register with $5 free credits, pay with USDC on Base, no human in the loop. Read the MCP guide →

Every API response is signed with ed25519 and timestamped. You can prove what was served and when, months later. That is what we mean by provenance-first.

Related reading: DePIN Intelligence Guide · DePIN Telemetry · How to Query DePIN Data · DePIN Data Verification · Pricing