Subnet News — September 21, 2026

SN88 Investing’s locked-alpha experiment gains a reported 9% allocation

Today’s SN88 Investing channel brief attributes a more specific design to the team: 9% of miner emissions for locked-alpha holders, with owner wallets excluded. This adds detail to the official “next week” experiment reported yesterday. It is a subnet-level incentive proposal, not a 9% share of all Bittensor emissions. The percentage is channel-digest-attributed; an activated payout implementation and actual recipient distributions were not independently confirmed.

Why 3 stars: A specific reward allocation is useful new information, while payout activation remains unverified.

-0.0% α/TAO 24H price action as of Sep 21, 2026, 10:28 AM ET

SN61 RedTeam ties its buyback restart to pending Kraken approval

SN61 RedTeam’s channel brief says the team expects approval of its Kraken buyback account next week and plans to resume buybacks afterward. It also describes a transparency website for buyback and conviction figures. On the previously discussed integration, the brief says contracting is expected by year-end and installations with three unnamed major banks are targeted for Q1 2027. Those remain future milestones attributed to the digest: this is not confirmation that buybacks have restarted or that three new bank contracts have been signed.

Why 3 stars: The dependency for restarting buybacks and the distinction between a commercial target and a signed deployment matter to holders.

-4.0% α/TAO 24H price action as of Sep 21, 2026, 10:28 AM ET

Emission watch: SN10 Pareton and SN71 Leadpoet return to zero allocation

At 10:28 AM ET, SN10 Pareton and SN71 Leadpoet show zero network allocation, versus approximately 0.000812% and 0.000239% respectively in yesterday’s saved snapshot. SN70 moves from zero to 0.008036%; its current metadata has no project name. These compare two observed snapshots, not an uninterrupted history or confirmed miner payouts. Network allocation and a subnet’s voluntary burn policy are separate. The data does not establish manual intervention by Const, Arbos or any other named person.

Why 3 stars: The zero/nonzero changes continue the requested emissions watch without overstating tiny allocations or inventing causes.

StonkFun adds AVAX pairs and reports $1.10M daily revenue

StonkFun says AVAX-paired launches are now available through Sunrise. Its September 20 report lists $1,100,719 revenue, $660,554 buybacks and 2.35 million STONK burned, while another announcement puts cumulative rewards to holders of reward coins at $65 million. These are platform-reported figures, not audited accounts or TAO-only revenue and burns. The new pairing is AVAX; previously covered TAO launch pairs are not being announced again.

Why 3 stars: A new supported pairing and fresh operating totals update the broader StonkFun ecosystem watch.

SN118 Ditto discusses stronger anti-abuse checks and a 5% development treasury

SN118 Ditto’s team-attributed brief outlines possible audit, quality-scoring and reviewer changes, plus a proposal to direct 5% of emissions to developer bounties with deposits for issue claims. These are design discussions, not a confirmed activated treasury or an implemented package of eight fixes. The beta activity log is already public and was covered previously.

Why 3 stars: A proposed change to evaluation controls and development funding deserves coverage before activation is assumed.

SN8 Vanta changes the pro-account transition to the weekly payout boundary

SN8 Vanta’s updated rules replace a fixed seven-day wind-down with a transition ending at the first Monday 00:00 UTC after entry, processed on the next refresh. During wind-down, traders can close or reduce positions but cannot expand them; promotion can still be requested earlier. Remaining positions close when the transition completes. This is a verified repository and documentation change; adoption by every live validator was not independently checked.

Why 3 stars: The timing change can materially shorten a trader’s transition and affect open-position handling.

SN11 TrajectoryRL releases a smaller evaluation suite and clearer validator health

SN11 TrajectoryRL’s 0.7.3 code prepares SPEC 26, reducing the suite from 26 scenarios to 20 after an analysis of 614 challenger packs identified tasks with little ranking signal relative to their cost. The release explicitly requires a paired server change and ordered rollout before activation, so the new maximum score should not be assumed live everywhere. A separate fix reports validator infrastructure failures in heartbeats and prevents a session that never ran properly from being posted as a miner’s zero score.

Why 3 stars: Evaluation cost, score comparability and distinguishing infrastructure failure from miner performance directly affect the competition.

SN15 ORO prepares six task families and re-queues qualifying agents

SN15 ORO’s team-attributed brief specifies six task families for a new qualifying suite after the current race, with greater separation between qualifying and race material. It says surviving agents will be re-queued and submission cooldowns cleared. This develops yesterday’s general suite plan with concrete scope and transition handling; completion of the rollout was not independently confirmed.

Why 3 stars: New task coverage and re-queue rules affect which agents qualify for the next race.

SN21 AdTAO makes admission verdicts identify their evaluation dataset

SN21 AdTAO’s admission pipeline now checks a served held-out dataset against its published hash and records the dataset identity in each verdict. If neither a served nor cached dataset is available, the public training bundle remains an explicitly labelled fallback. This lets observers tell whether two verdicts used the same evaluation material. The held-out examples themselves are not made public, and the change does not establish a new model-performance result.

Why 3 stars: Identifying evaluation data and fallback use makes admission decisions more comparable and auditable.

SN34 BitMind refreshes audio holdouts and re-evaluates the incumbent

SN34 BitMind’s channel brief reports refreshed audio holdouts and re-evaluation of the current king. The team is considering smaller iterative benchmark changes that avoid full resets and an audio augmentation metric. The fresh holdout work is distinct from yesterday’s reported benchmark reset; the proposed metric and future reset policy should not be read as already deployed.

Why 3 stars: Updated hidden tests can change whether an incumbent’s performance generalizes.

SN36 Epago replaces one-shot entry with three attempts per hotkey

SN36 Epago’s new competition rules allow each hotkey up to three attempts counted from round four, with one model per hotkey per round. A newer accepted reveal replaces a waiting model without consuming an attempt; an attempt is charged when the round takes the model into its field. Cooldowns are removed, and a narrowly defined near-miss can re-enter once within the same three-attempt allowance. This is a code-and-rules update, not proof of a new winning model.

Why 3 stars: The new entry rules reduce the cost of stale queued submissions and give miners a clearer retry path.

SN44 Score adds a petrol-station detection task amid server migration

SN44 Score’s channel brief reports a new petrol-station detection element covering canopies, price boards and hoses. It also reports server migration affecting challenge delivery and a registry-filter problem around block 9,112,200. The useful fresh development is the specific vision task and operational disruption. Recirculating Score Studio launch stories and customer speculation do not establish a new signed petrol-station contract or a disclosed revenue figure.

Why 3 stars: A new applied vision task and interruption to challenge delivery are relevant to both product watchers and miners.

SN51 Lium lowers the idle-payment cap for single-card B300 capacity

SN51 Lium’s merged incentive configuration reduces the B300 single-card idle bucket from ten cards to four, while keeping the eight-card bucket at 32 GPUs. The change cites observed renter demand and excess idle supply. Free cards in partially rented split machines also share the single-card bucket when their split minimum is one. This affects idle-capacity incentives; it is not a cut to all rented B300 payments or evidence of lower customer demand across every GPU type.

Why 3 stars: A concrete payment-cap change matters to compute suppliers and how emissions subsidize unused capacity.

SN56 Gradients plans a check for exposure to Intercode test material

SN56 Gradients’ team-attributed brief says it is addressing training setups that expose Intercode evaluation material and plans an automated check, with manual submission review in the meantime. The team says tournament rules will not change before further discussion with miners. The operational issue is evaluation integrity; no individual miner allegation or completed automated fix is asserted here.

Why 3 stars: Protecting hidden evaluation data is material to the fairness of model rankings.

SN71 Leadpoet prepares longer Arena runs and previews integrated outreach

SN71 Leadpoet’s repository adds a 60-minute execution profile for new Arena rounds and a guarded migration for the unstarted September 22 round, preserving historical 45-minute results. Separately, founder Gavin Zaentz says LinkedIn and email outreach will soon connect to the same system that finds buying signals and contacts. The longer profile is implemented in code; the outreach workflow is an upcoming product feature, not a confirmed launch.

Why 3 stars: More execution time changes miner preparation, while the announced outreach direction explains the intended customer workflow.

SN80 OpenRoboto prioritizes new benchmark and data tracks

SN80 OpenRoboto’s latest channel brief says new benchmark and data tracks are expected next week, with emission shares to be adjusted. It also says a physical-robot track is not the immediate priority and no common robot standard has been finalized. This is a roadmap update following yesterday’s simulation comparison; it does not announce live new tracks, a confirmed allocation split or a physical deployment.

Why 3 stars: The next evaluation priorities matter to miners choosing where to build, with timing and allocation still pending.

SN104 TaoStatus moves reward calculation to the previous completed epoch

SN104 TaoStatus’s merged reward code pays each epoch using verified usage reports from the preceding completed epoch, replacing the carried-forward moving average. A miner with no qualifying report in that window gets no weight from this calculation, and an accumulator that spans a multi-epoch validator outage is discarded. The documented burn fallback still applies when the usage pipeline is unconfigured. This confirms new reward logic in the repository, not universal live adoption or a new network emission allocation.

Why 3 stars: Reward freshness and removal of stale carried-forward scores materially change miner incentives.

SN120 Affine’s team reports Arbos leading with an automated training script

SN120 Affine’s channel brief attributes to Const an account of team-controlled Arbos leading the competition using an automated GRPO training script costing about $500 per day in compute. It says logs and scripts were shared and calls for more competing miner teams. The figure and ranking are team-reported, not independently reproduced here. This describes competition participation; it is not evidence that Const or Arbos manually switched network emissions on or off.

Why 3 stars: Team participation and the claimed training cost are relevant competitive context, while the emissions interpretation must remain evidence-based.