Originally published at mrnasdog.com/research/tao/inflation by MrNasdog.
Bittensor minted about 324,000 TAO over the last 90 days — the block reward, and the only thing on this chain that creates supply. Registration and network fees recycled only ~12,600 TAO back into the unissued pool, under 4% of what was minted, so the MrNasdog Pressure Framework reads TAO at about +3.24% net. That figure divides real on-chain issuance by a supply feed that has been frozen at 9,597,491 TAO since Aug 17 2025; measured against the chain's own ~11.19M issued, the same flow is about +2.79%. TAO is a fair-launched, hard-capped asset that still dilutes at roughly the pace of pre-halving Bitcoin — the ceiling is real at 21,000,000, but 9.81M TAO of it is still unissued.
The verdict, in one paragraph
For the 90 days to Aug 3 2026, the MrNasdog Pressure Framework reads TAO at about +3.24% net: sell pressure of 324,000 TAO against buy pressure of ~12,600 TAO. Our supply monitor reads -0.01% for the same period — a gap of 3.26 percentage points, large enough that a monitor-gap flag ships with this page. The gap is not a disagreement about Bittensor's mechanics; both readings agree the chain mints 3,600 TAO a day. It is a stale denominator. The classified float figure our monitor consumes has sat frozen at 9,597,491 TAO since Aug 17 2025 — eleven months of a constant — while the chain's own explorer and independent market trackers both read about 11.19M TAO issued. Because dividing real issuance by that frozen feed produces a higher percentage than the chain's live base, and because the two denominators fall on opposite sides of a score band, this page ships the frozen-feed reading and flags the on-chain one for review. TAO is best labelled a hard-capped chain that is still early in its emission curve: scarce at the ceiling, inflationary today.
Sell pressure: where new TAO comes from
From one place only. Sell #1 — protocol inflation — is 324,000 TAO, and it is the entire sell ledger. Bittensor pays a fixed reward on every block, and that reward is a step function of how much TAO has been issued: 1.0 TAO per block until 10,500,000 had been issued, 0.5 TAO until 15,750,000, then 0.25. The first Bittensor halving fired when issuance crossed 10,500,000 on Dec 15 2025, cutting emission from 7,200 TAO to 3,600 TAO a day. Since then the subsidy has been 0.5 TAO per block at Bittensor's twelve-second blocks, which over roughly 648,000 blocks in a 90-day window is 324,000 TAO of gross emission. Issuance is nowhere near the second halving mark — at the current pace 15,750,000 is roughly 1,300 days away, around 2029.
Every other sell row is zero, and for TAO those zeros are structural rather than lucky. Sell #2 — vesting unlocks — is zero because Bittensor had no presale, no ICO and no private round. There is no team allocation, no investor allocation and no foundation allocation, therefore no vesting contract and no unlock calendar; every TAO in existence was mined. Sell #3 — Foundation and unscheduled unlocks — is zero: the Opentensor Foundation holds only a small operating balance and no scheduled release, and the large institutional holders — venture firms such as Polychain and Digital Currency Group — bought their TAO on the open market and by validating, not from a locked tranche that could dump. Sell #4 — long-term locked or bankruptcy — is zero, because no estate, trustee distribution or expiring lock-up contract touches TAO. The only supply overhang Bittensor has is the 9.81M TAO that has not been issued yet, and nobody can hand that out early — it comes out through the block reward already counted in Sell #1.
Buy pressure: where new TAO goes
Buy #2 — protocol fee burn — is ~12,600 TAO, and it is the only offset TAO has. Bittensor does not burn in the usual sense; it recycles. TAO spent registering a subnet, registering a miner or validator slot, or paying network fees is returned to the unissued pool rather than destroyed, which removes it from the market today and pushes the next halving date further out. Against gross emission of 324,000 TAO, a recycle of ~12,600 TAO leaves net new issuance of about 311,000 TAO — the recycle offset under 4% of the mint this quarter.
The important property of the Bittensor recycle is that it is lumpy, not smooth. Registration demand arrives in bursts: a fortnight earlier in 2026 recycled almost 39,000 TAO, while the stretch since May 2026 has recycled almost nothing, and this quarter caught the quiet part — which is exactly why the offset is so small. The remaining buy rows are all zero. Buy #1 — programmatic buyback — is zero because Bittensor operates no buyback contract and points no revenue stream at one. Buy #3 — Foundation buy — is zero because the Opentensor Foundation holds no allocation to deploy and has disclosed no purchase programme. Buy #4 — new long-term lock — is zero: about 67% of issued supply sits staked, but Bittensor staking can be undone at will, so it is custody rather than escrow. The Conviction time-lock that shipped with the v431 network upgrade on Jul 18 2026 locks subnet alpha tokens for ownership contests, not TAO.
Foundation and overhang
TAO carries one of the smallest team-controlled overhangs in the catalogue, and that is a design property rather than a phase. Bittensor fair-launched in 2021 with no premine and no ICO, so there is no founder cap table, no investor tranche and no DAO treasury holding TAO. The single enumerable item is the Opentensor Foundation's own operating balance, a rounding error against 11.19M issued, with no scheduled release. The Foundation also runs a validator that third parties delegate to, but that stake belongs to the delegators and can be withdrawn by them, so it is not foundation-controlled supply. The one genuinely large pool is the 9.81M TAO that remains unissued, and it is governed by protocol code rather than by anyone's discretion. If the Foundation balance falls between refreshes, that outflow enters Sell #3 at the next refresh; on current size it would take a change of several orders of magnitude to matter.
How TAO compares to other hard-capped halving chains
The natural peer group is hard-capped, halving-model chains, and Bittensor is a deliberate copy of that design: 21,000,000 maximum supply, a subsidy that halves at fixed issuance milestones, and a fair launch with no allocation to insiders. On the scarcity axis TAO is genuinely in that family, and it is a much smaller family than the market implies — most decentralised-AI tokens are uncapped, pre-allocated, or both. Where TAO differs from Bitcoin is position on the curve. Bitcoin is roughly 95% issued and dilutes well under 1% a year; Bittensor is 53% issued and dilutes at a few percent a quarter. TAO today looks like Bitcoin around its second halving, not Bitcoin now.
Against uncapped continuous-emission layer 1s that mint new supply every block to pay validators, TAO's advantage is the ceiling: its dilution is finite and pre-committed, and no governance vote can raise the cap or change the schedule. Against exchange tokens that run quarterly buyback-and-burn programmes, TAO is the weaker structure, because it has no revenue-funded bid at all — the recycle is a by-product of registration demand, not a policy, and at ~12,600 TAO it offset under 4% of emission this quarter. And against fee-burn chains, the Bittensor recycle is a softer mechanism by construction: recycled TAO is not destroyed, it is queued for re-emission later, so the effect is to delay dilution rather than cancel it.
What to watch in the next 90 days
The first item is whether the frozen supply feed is repaired — the classified float has read 9,597,491 TAO since Aug 17 2025 while the chain issued another 1.6M, and any correction snaps the headline percentage toward the on-chain +2.79%. The second is subnet registration demand, which is the entire Buy #2 row — a return to the earlier pace would multiply the offset several times over without any protocol change. The third is the US securities decision window on the Grayscale and Bitwise spot TAO products, tracked for Aug 2026; approval would not remove a single TAO from supply, but it would change who holds the float. The fourth is the emission-distribution model, which reverted to price-based weighting in Jun 2026 after seven months on a flow-based model; it changes which subnet is paid, never how much the chain mints, and any proposal that changed the latter would be the single most important event this ledger could record. The fifth is the second halving mark at 15,750,000 issued, roughly 1,300 days out — far outside this window, but the date moves with every quarter of recycle.
Summary
TAO is a fair-launched, hard-capped token with exactly one source of new supply and almost nothing pushing back. Over the last 90 days Bittensor minted 324,000 TAO at 0.5 TAO per block while registration and fee recycling returned only ~12,600 TAO to the unissued pool, leaving the framework at about +3.24% net on the frozen reference feed, or +2.79% on the chain's real ~11.19M issued base — against a monitor reading of -0.01% that traces entirely to a float figure frozen since Aug 17 2025. The key risk is not a hidden unlock, because there is none: it is that the only offset TAO has is a by-product of network activity, so dilution runs at its full rate whenever the network is quiet. The ceiling is hard at 21,000,000 and cannot rise, but 9.81M TAO of it is still to come.
MrNasdog Pressure Framework analysis of TAO, Metric 1 — Inflation. Data + explanation only. Not financial advice. Updated Aug 3 2026.
Top comments (0)