Source and exclusions
We read Taostats’ Bittensor event index for AlphaBurned and AlphaRecycled outcomes. We convert the emitted integer amount from base units by dividing by 10⁹. This includes combined buy-and-burn operations and burns executed inside other calls. Each event is counted once; AddStakeBurn is supporting evidence, not an additional burn. Requested amounts and TAO purchase inputs are not substituted for actual alpha burned. Burn and recycle event types remain identified in the evidence. The index is not an independent guarantee of complete chain coverage.
Burns, buybacks and revenue
A successful burn call establishes an on-chain action. It does not establish that the tokens were purchased with revenue, identify a treasury, or prove a business is profitable. Buying and holding alpha leaves no burn. The claim register quotes SubConnect’s Bittensor Revenue Index V2 from August 2026, reviewed on 3 September 2026; it is dated third-party reporting, not our verification of funding.
Verified buy-and-burn operations
A separate collector scans AddStakeBurn events across all subnets over a rolling 30-day window. Each receipt must match StakeAdded, AlphaBurned and a successful dispatch in the same transaction, with matching participants, subnet and amounts. The TAO figure is the emitted purchase input; the alpha figure is the matched burn outcome. Receipt IDs preserve distinct operations inside batches. These receipts are displayed separately from tracked-wallet trades and are not added to burn values or purchase charts. The sources can overlap. Funding and team attribution remain unverified. Collection runs hourly after backfill, with bounded five-minute retries during incomplete scans. The scan cutoff and incomplete/stale status are shown. A receipt can be newer than the daily burn index. Only receipts from the known current subnet registration are displayed.
Valuation and time windows
The rankings multiply alpha burned in rolling 24-hour, 7-day and 30-day windows by the latest daily panel price in TAO. These are comparable current-price values, not historical expenditure. Subnet charts show burns in TAO valued at each UTC day’s panel price, alongside tracked-wallet TAO spent on buybacks. History CSVs also show alpha amounts and daily TAO values, with a blank when a needed price is missing. Today’s calendar day is partial. Seven-day activity change compares alpha with the preceding seven days; a zero baseline has no percentage change.
Freshness and coverage
Collection runs daily. “Burn collection” is the start time of the most recent scan that completed both call types; it is independent of the latest observed burn. We flag collection older than 36 hours, and say “not recorded” until a completed scan is recorded. No indexed records on a day means no matching records in our database, not proof of complete coverage. Historical backfill coverage has not been independently certified.
Subnet identities
The directory lists subnets in the latest panel. History follows the subnet number, which may have been reused after deregistration; it must not automatically be attributed to the current team. A claim retains the name and date recorded by its source.
Revenue to alpha: how we classify a buyback claim
“Buyback” is one word for at least four different on-chain actions: alpha bought from the pool and burned, alpha bought and held, alpha withheld from emissions and burned, and alpha recycled back into the emission pool. Each has a different supply effect, and each leaves a different trace. A claim cannot be checked until it is split into three questions: how the alpha was acquired, what happened to it, and where the funding came from. We classify every claim on those three axes and publish the answer beside the figure.
Axis A: acquisition
| Code | What we observe | How we verify it |
|---|---|---|
| A1 Pool purchase | TAO in, alpha out, from an attributed wallet | Indexed trades for the wallet. Verified only when the wallet is on the reviewed registry or is the current owner coldkey. |
| A2 Purchase inside a buy-and-burn call | StakeAdded and AlphaBurned in one transaction, matched on participants, subnet and amounts | The AddStakeBurn receipt. The TAO input and the alpha burned are both emitted outcomes. |
| A3 Emission-retained alpha | Miner incentive directed to an owner-controlled hotkey, then burned or recycled | Excluded from buyback figures. It was never bought. Counted in burn totals when the burn event exists, and labelled. |
| A4 Not observed | No attributed wallet, or an attributed wallet with no trades in the window | Reported as “no on-chain evidence”, never as zero purchases. |
Axis B: disposition
| Code | What we observe | Supply effect |
|---|---|---|
| B1 Burned | AlphaBurned event | Removed from circulation. |
| B2 Recycled | AlphaRecycled event | Returned to the emission pool. Not equivalent to a burn. |
| B3 Held or staked | Alpha remains in an attributed wallet after purchase | No supply change. Visible only if the wallet is published. |
| B4 Sold back | Purchase followed by a sale from the same wallet | Nets toward zero. We report gross and net. |
| B5 Distributed | Alpha moved to stakers, miners or a third party | Not measured today. Shown as “not classified”. |
Axis C: funding
Nothing on the chain says where TAO came from. A wallet that buys alpha may be funded by invoices, by emissions, by a treasury sale or by a founder. So this axis is never marked verified. It records the strongest public statement and its date.
| Code | Source |
|---|---|
| C1 Team statement with receipts | The team publishes revenue and the purchase transactions, or a dashboard that links them |
| C2 Team statement without receipts | A policy statement, for example “100% of revenue to buybacks”, with no linked transactions |
| C3 Third-party report | A report such as the SubConnect Bittensor Revenue Index, quoted with its date |
| C4 No statement | Nothing found by the source monitor or submitted by the team |
A fully verified row is A1 or A2 acquisition, B1 disposition, C1 funding. The chain proves the purchase and the burn. The team’s receipts connect them to revenue. We can currently verify A and B. We can only cite C.
The two ratios, and why we do not publish them yet
Comparing one subnet’s daily emissions with an annual revenue estimate mixes different units, different windows and different sources. The comparison that can be made from the chain is the supply ratio, alpha burned (B1) divided by alpha emitted to that subnet in the same UTC window, and the buy-pressure ratio, TAO spent on attributed purchases (A1 plus A2) divided by TAO injected into that subnet’s pool in the same window.
Both numerators exist in our data today. Both denominators need a validated per-subnet emission series in matching units, read from finalized blocks, not from a dashboard’s daily figure. Until that series passes the same checks as the burn index, the ratios stay unpublished. A burn figure on its own does not establish net supply contraction.
What we ask a team to publish
Four things let us move a row from “no on-chain evidence” to a verified A and B:
- The wallet or wallets the programme buys from, as SS58 addresses on a page you control.
- The rule: what share of what revenue, on what cadence.
- The disposition: burn, recycle, hold, or distribute.
- A page we can monitor for changes. The source monitor re-reads it daily.
Submit at /subnets/submit-wallet. We attribute within a day of review and say publicly that we did.
Check a figure
Subnet pages link to the indexed transactions. You can also read the Taostats event API documentation. Report a discrepancy to hello@subnetmetrics.com, including the subnet number, window and transaction hash.