September 5, 2026 · Published daily
Subnet News — September 5, 2026
Yanez’s enterprise disclosure leads at five stars with a 20.55% α/TAO move. Score’s Shell and Eni rollout announcement gets a clearly dated five-star catch-up. Sentios, Ridges, MicroTensor and major competition changes follow in today’s ranked edition, researched through 9:32 AM ET.
- Coverage window
- Sep 4, 2026, 2:37 PM – 9:32 AM ET
- Subnets in coverage roster
- 128
- Subnets with updates
- 46
Biggest stories first
Today's ranked news
Yanez names Dow Jones, Moody’s and LexisNexis among customers and partners
Yanez publicly disclosed customers and partners on its website, including Dow Jones, Moody’s and LexisNexis. Its September 4 update also reports successful initial proof-of-humanhood and uniqueness deployment across hundreds of devices. The disclosure does not specify which relationships are customers versus partners, contract values, or which product each organization uses. SN54 gained 20.55% over 24 hours in the 9:32 AM ET α/TAO snapshot.
+20.5% α/TAO 24H price action as of Sep 5, 2026, 9:32 AM ET
Catch-up: Score founder adds Shell and Eni stations to rollout
September 3 announcement, added to our coverage today: Score founder Max Sebti said Shell and Eni stations were added to the rollout. This is a significant real-world deployment lead for SN44, but the announcement gives no station count, contract value or geographic scope. Separate confirmation from the two brands was not located; it should not be read as a rollout across their entire global networks.
Ridges v0.3.0 adds a framework for multiple agent competitions
Ridges merged v0.3.0 with separate competition policies, lifecycle management, emission weights, submissions and leaderboards. The release provides the foundation for agents to compete across more than one contest. This verifies the software change; it does not establish that every planned competition is already open or generating customer revenue.
+5.3% α/TAO 24H price action as of Sep 5, 2026, 9:32 AM ET
MVTRX updates its trading interface; new scoring remains under review
The team reports a redesigned swap page, better chart controls and an expanded Hangar for local testing ahead of miner grants. Version 0.6.1’s earnings-sensitive scoring change is still being tested and has not been confirmed live. The α/TAO market snapshot showed a 27.65% 24-hour decline, making the rollout and pending scoring decision material to watch.
-27.7% α/TAO 24H price action as of Sep 5, 2026, 9:32 AM ET
MicroTensor brings the classification arena into its live competition setup
New commits add label-accuracy scoring for the classification arena and configure classification as the live track, alongside coordinator and round-window fixes. The team’s September 5 brief describes Arena 4, intent classification and resource limits, while acknowledging deployment-configuration issues. SN92 was up 24.38% over 24 hours in the 9:32 AM ET α/TAO snapshot.
+24.4% α/TAO 24H price action as of Sep 5, 2026, 9:32 AM ET
Catch-up: Yuma reports 1 million TAO staked through its validators
Yuma announced the milestone on September 1, according to TAO Daily’s September 4 coverage. The reported total is TAO staked through validators running on Yuma-owned hardware; it should not be described as a million-TAO corporate treasury or custodial balance. This edition did not independently reconstruct the validator total.
TAOApp Wallet beta brings hardware signing, proxies and multisig into one interface
Fresh September 4 coverage highlights TAOApp Wallet’s browser beta and its support for watch-only accounts, Ledger and offline signing, proxies and multisig. The Chrome listing shows version 0.5.0 updated September 2. This is a catch-up on an existing release, not a new wallet launch on September 5; the listed security features are not a substitute for an independent security audit.
DSperse investigates zero-weight epochs and a reported 94.66% effective burn
The team is working to align emissions with leaderboard results and investigate epochs where completed work does not persist in weights. A community calculation summarized in the September 5 brief puts average effective burn at 94.66% across 20 epochs. That figure has not been independently reproduced. The team recommends version 14.14.2.
BitMind targets Sunday for v5 and an 85/10/5 King of the Hill split
BitMind’s team says v5.0.0 is scheduled for Sunday, September 6. The proposed reward split gives each modality’s current winner 85%, then 10% and 5% to its two predecessors, with one submission per hotkey. A new public commit burns discriminator incentives while waiting for King of the Hill. Release activation and resumed payouts were not confirmed by this edition’s cutoff.
Bitsec’s Sentios offers continuous security reviews from $1,000 a month
Sentios offers ongoing adversarial code review, checks as code changes, and reproducible findings with GitHub and Slack integrations. Its live pricing starts at $1,000 per month for continuous coverage; one-off scans are also offered. TAO Daily’s September 4 coverage identifies the subnet behind Sentios as Bitsec, SN60, despite incorrectly naming Bitcast in its headline. Customer outcomes and revenue were not independently verified.
Catch-up: Chutes highlights a year of production traffic for serving research
In its September 1 update, Chutes highlighted research covering 6.12 billion requests and 35.8 trillion input tokens, co-authored with researchers at Harvard and UChicago. The linked paper studies workload evolution, caching and load balancing using a year of production traffic. Chutes says the trace contains request metadata rather than user prompts. This is a late-found research story, not a September 5 release.
Leadpoet completes an end-to-end miner model submission test
A September 5 validation record documents CLI upload, sandbox execution, judging and a persisted public result on isolated testnet SN401. Rewards were disabled and production was unchanged; the tested miner scored 0 against a 1.02 baseline. Separately, the team’s current discussion reports inactive fulfillment and delayed Leg 1 grading while infrastructure repairs continue.
Vidaio calls for miner v0.5.4 during infrastructure migration
The team’s September 5 brief says infrastructure migration is around 70% complete and miners must move to v0.5.4 using the new CPU-based onboarding guide. Validators are directed to a thin-validator setup without a GPU requirement. Temporary disruption is possible during migration; the first compression competition is planned for next week, with details still pending.
Cortex adds its own inference judging backend and narrows live challenges
Public commits add an InferenceOffer backend for RLM judging and remove Prism, Relearn and Design from the active product challenge set, retaining Bounty and Proof. The changes focus the network on code bounties and research proofs. They verify a change in the software and evaluation structure, without establishing new customers or revenue.
Capcomp reports its first cycle of sealed on-chain mining commitments
The team reports 27 recipes sealed before results existed, using timelock encryption and drand randomness, with run bundles on Hugging Face and digests committed on-chain. It plans to expand the adapter pool from 30 to more than 50. The repository documents the commitment workflow, but this edition did not independently replay the 27 recipes or audit their on-chain records.
Minos reports reproducible scores; an emissions restart is still unconfirmed
The team says shared ground-truth files now make scores reproducible. The same discussion includes a request to re-enable emissions, but no confirmation that the restart occurred. Although the secondary brief’s headline says emissions resumed, its underlying summary supports only a request. Validator ranking and reward mechanics remain under discussion.
Ditto ships coding practice tasks and encrypted benchmark delivery
Ditto added ten public coding practice tasks and encrypted private benchmark payload delivery through Hippius. The latest v0.217.0 release also follows a hosted control admission API that is off by default. These are inspectable preparation and security changes for the coding track; they do not establish a generally available hosted coding launch.
Affine makes submissions private; private-duel issues are under investigation
Catch-up from September 4: Affine’s official account says submissions are now private, crowned models remain public, and losing checkpoints will be released later for verdict review. The September 5 team discussion also acknowledges a private-model evaluation problem and reviews overfitting penalties. The privacy change should not be mistaken for confirmation that private dueling is currently working reliably.
The DeFi Report publishes a new Bittensor research profile
Michael Nadeau’s September 4 Watch List examines Bittensor’s economics, activity and investment thesis. It provides an outside research perspective rather than a protocol release or a verified forecast. Readers can review the original analysis and its assumptions directly.
Catch-up: Root claim costs draw scrutiny over broad subnet allocation
TAO Daily’s September 3 evening report summarizes Mentat Minds’ concern that claim costs rise as validator baskets include more subnets, potentially favoring concentration. This is analysis of the existing claiming mechanism, not an announced protocol change. Proposed remedies and the size of the economic effect still require evaluation.
Teutonic resumes evaluations and reports faster processing
The team reports evaluations have resumed after maintenance, with provider download issues addressed and evaluation time reduced from roughly 6,000 to 1,800 seconds. Model digest visibility was added to the queue. On-chain hash commitments were discussed, while the team continues to rely on post-commit manifests for integrity checks.
Allways 3.3.2 adds settlement safeguards for destination contracts and addresses
The 3.3.2 release follows fixes that reject malformed destination addresses during attestation and settlement and apply a gas floor to contract destinations. These are concrete reliability improvements for cross-chain swaps. Allways Access also provides a documented browser-based route into the service; this edition does not treat that existing interface as a new launch today.
Vanta fixes equity recovery as the team answers emissions questions
A new commit restores current balances and live equity when rebuilding from a snapshot. Separately, Vanta says trader payouts exceeded emissions threefold last week and were supplemented by the team following an emissions cut. The accounting fix is public; the payout ratio and funding claim come from the team discussion and have not been independently reconciled.
Pareton proposes private submissions and fixed rounds to address spam
Pareton’s team is considering one- to two-day private submission periods and fixed 12- to 24-hour rounds after an exploit disrupted participation. These are proposed mechanism changes rather than confirmed deployed rules. A final implementation and activation time were not available in the reviewed discussion.
Compute Horde clarifies transfer restrictions and flags fake support accounts
The team says its subnet token is non-transferable and distinguishes unstaking from transferring stake on transfer-disabled subnets. It also warns about impersonated support accounts and unsolicited wallet instructions. Users encountering withdrawal problems should consult the project’s official guidance; this report does not infer a regulatory status from transfer restrictions.
Cacheon addresses baseline drift ahead of a planned Monday arena
The team acknowledges inaccurate dashboard speed readings and emission discrepancies tied to relative baseline improvements. Validator restart and baseline adjustments are underway. Its next GLM 5.3 benchmarking arena is planned for Monday, September 7; current dashboard readings and a scheduled arena should not be confused with verified final rewards or a completed launch.
AdTAO moves incentive amendments to September 5 and fixes top-20 rewards
A public amendment changes the effective date to September 5. A related fix makes the top 20 miners by standing eligible for the reward curve even when relative performance is zero. The commits provide concrete evidence of the rule change and implementation; chain-level reward results were not independently replayed.
Trishool follow-up: scoring retries can stretch beyond a day
Following yesterday’s five-star Trusted Access for Cyber announcement, Trishool’s team describes a 370-second validator timeout and repeated retries that can leave scoring incomplete for 24 hours or more. This is an operational follow-up. The OpenAI program acceptance remains the September 4 story and is not evidence of a broader commercial partnership with OpenAI.
GM’s pricing explainer details how inference suppliers compete on cost
New September 4 coverage explains SN28’s model: an OpenAI-compatible gateway routes requests to independent miners that set prices and bring their own provider access. Confidential execution supports the delivery model. This adds commercial context to the model-access news already covered yesterday; advertised discounts are variable and were not independently benchmarked in this scan.
Score operators report download bottlenecks and validator inconsistencies
Separate from Score’s commercial rollout lead, the team discussion describes persistent challenge-video download delays exceeding 380 seconds and inconsistent validator trust readings. A reported server fix had not resolved the issue for all affected miners. The scope and effect on final rewards still need clarification.
Feval improves evaluations and says its live dashboard is online
Public commits refine evaluation, dataset handling and verification. The team says its dashboard now shows live evaluation data, while additional validator participation and stake allocation are still pending. Prospective Kraken allocation and Yuma participation should not be reported as completed integrations.
Engy seeks built-in support in opencode and deepseek-harness
Engy’s team is requesting approval to become a built-in provider in opencode and deepseek-harness. The reviewed discussion confirms the request, but does not establish acceptance, a merged integration or availability to users.
Gradients limits tournament entries by coldkey
A new change enforces one tournament entry per coldkey, preventing duplicate participation through multiple hotkeys. It also deduplicates retry handling and rejects a second fee enrollment. The constraint applies at the controlling wallet level, which is more restrictive than one entry per hotkey.
RedTeam remains active while Docker build failures stall submissions
The team confirms the subnet is operational, but some miners report submissions stuck at Received for more than 26 hours. The discussion attributes these cases to Docker container build errors during scoring. That identifies a submission-processing problem rather than confirming a subnet-wide shutdown.
Gittensor’s SparkInfer 0.5.4 adds image input and reproducible benchmarks
SparkInfer 0.5.4 adds image-input support and longer-context DSpark benchmark coverage. The team also committed the workload corpus and measurement harness after acknowledging that previous published benchmark numbers could not be reproduced from the repository alone. This is a new verification and capability update following yesterday’s compute-serving launch.
Hippius bridge update awaits runtime rollout after a merged fix
The team points to Subtensor PR 3140 for a forthcoming bridge-related update. The fix, merged September 4, restores contract-proxied stake transfers affected by a runtime filter. Merge status is confirmed, but runtime activation and restored Hippius bridge operation were not confirmed by this edition’s cutoff.
UMI follow-up: reference implementation is available, translation rewards are inactive
Following SN78’s pivot from Vocence, UMI has published its reference implementation, baseline model and inspectable miner pilot, including Apple Silicon support. The repository explicitly says public calibration has not started and translation weights remain inactive. Component review and offline rehearsal are open; this is not confirmation that the new translation competition is paying rewards.
OpenRoboto clarifies the season’s grasp-and-lift baseline
A September 5 documentation change specifies a green-block grasp-and-lift baseline, including object dimensions and the requirement to lift clear of the table. Placement is not required for this baseline. The locked season task specification remains authoritative, giving miners a more precise target for local testing and submissions.
Beam reports petabyte-scale transfers and hardens encrypted room handoffs
Beam’s September 4 official update reports petabyte-scale data flowing among services including Hugging Face, Hippius and major object stores. That volume is a project claim without an independently verified measurement window. A September 5 commit separately requires end-to-end-encrypted room-transfer capability during handoffs.
Nodexo explains executor-related 503 errors and stale canary flags
The team attributes affected validator API 503 responses to missing executor registry entries. It also says flags left by old IP and port configurations clear after a successful canary check, rather than simply with time. The discussion does not support describing the issue as a system-wide outage.
Finsight reports a working testnet prototype and targets integration within two weeks
Finsight’s developers say a testnet prototype is working and Bittensor integration testing is underway. They target completion within two weeks and plan to share the repository. This remains a team-reported development milestone; production launch and the timetable have not been independently confirmed.
Green Compute plans on-chain miner payouts for greater transparency
Subnet lead Josh says miner payouts will move on-chain so the community can inspect them publicly. The team is also working on additional transparency measures. The reviewed discussion contains the commitment, but no completed migration, activation date or reconciled payout history.
Sundae Bar’s SBC25 submission window closes Sunday at 2 PM EDT
The team says SBC25 is open from September 4 at 2 PM EDT until September 6 at 2 PM EDT. It closed the previous challenge early after reporting that its production-performance criteria had been met. The new deadline is actionable; the claimed enterprise performance has not been independently benchmarked.
CookingTAO outlines one-click mining beta and testnet progress
CookingTAO’s September 4 update describes a one-click mining beta, a developer management portal around 80% complete, and SN122 testnet work. These are separate stages of development; the portal is not reported finished and the update does not establish a completed mainnet product launch.
Swarm reports a required anti-copying update
The team says it deployed a required update intended to discourage copying. The reviewed discussion does not publish enough technical detail or performance data to assess its effectiveness. Operators should follow the official release guidance rather than infer new scoring rules from the short announcement.
404-GEN corrects warmup replacement-budget display
A new public commit fixes how the warmup replacement budget is presented. This is a small miner-interface correction, not a new generation model or a change to the underlying product offering.
UR continues settlement qualification and replay checks
A September 5 qualification checkpoint records partial settlement-test repairs and improved validation evidence. Broader replay, full qualification and integration remain pending, and the final live campaign has not started. This is a documented readiness update rather than confirmation of a completed production rollout.
Enigma improves its local challenge and benchmark workbench
A new commit improves the workbench for local challenges and benchmarking, giving contributors a clearer rehearsal workflow. This is a tooling update; it does not demonstrate a new cryptographic breakthrough or completion of the previously reported RSA challenge.
Compelle’s King of the Hill streak reset highlights the Elo-gap rule
The team discussion describes an 18-block streak resetting when the miner’s Elo lead fell below the required 200 points at the validator tempo. Operators also questioned eligibility and strategy classification. Those disputes remain unresolved; this edition does not treat copying or manipulation allegations as established facts.
Albedo reports a new benchmark winner while reviewing training variance
The latest team discussion reports that King CXVII beat Genesis on both cited benchmarks. It also highlights substantial variation between training runs and duels. The comparison is team-reported and has not been independently reproduced in this edition.
ConnitoAI reports restored validator status and repaired cohort grouping
The team says validator status is back online and the cohort-grouping problem has been resolved. This is a reported operational recovery; the scan did not independently replay resulting scores or reward allocations.
Audit trail
Coverage notes
Searched
- All 128 active subnet slots: current chain identities and all 128 SubnetRadar profiles.
- 100 dated September 4–5 SubnetRadar briefs, including the ecosystem channel; substantive claims cross-checked where public evidence was available.
- 148 public repository paths, including current and legacy mappings and the core Subtensor repository; 143 accessible, with paginated commit scans through September 5 at 9:32 AM ET.
- 128 dated indexed X searches by subnet name, with SN numbers included in later query batches; public project and profile feeds plus original X embed retrieval for surfaced posts.
- Signed-in AlphaGap feed, market signals and Yanez detail page; its 48-hour summaries checked for old information, scope errors and unit mismatches.
- TAO Daily, TAO Media, public Bittensor news and partner searches; Yanez customer page, Score founder post, Sentios pricing, Chrome Web Store and the original Chutes research paper.
- Shizzy Unchained YouTube Videos and Live tabs; latest IDs and durations checked. No newer upload or stream was found.
- Final chain-market snapshot at September 5, 9:32 AM ET. Displayed price moves are rolling 24-hour alpha prices relative to TAO, not USD returns.
Public-data gaps
- Signed-in X search redirected to login, so a complete per-account X timeline scan and authenticated Grok research were unavailable. Indexed searches produced no relevant current results; this is an access/indexing gap, not evidence that every subnet was quiet.
- Public feeds and embeds recover some X stories, but private, deleted, unindexed and video-only details can still be missed. Partner-side confirmation of the Score rollout was not located.
- Discord discussions were reviewed through dated AI-generated briefs rather than authenticated original message history. Uncorroborated operational reports and metrics remain developing; relative dates were not treated as exact launch dates.
- Five mapped repository paths returned 404: deprecated/deprecated, actual-computer/actual-subnet-95, Satori119/Satori, NIOME-BITTENSOR/NIOME and Oh-My-Ditto/ditto. Alternate accessible paths were scanned where found; organization pages are not complete repository inventories.
- Catch-up items identify earlier event dates. Previously covered Chainlink/Forever Money and Trishool announcements are not presented as new September 5 announcements. Earlier dated editions remain unchanged.
- No independently verified forward alpha-unlock calendar was available, so Unlock Watch is empty. Research coverage ends at 9:32 AM ET; later announcements fall into the next coverage window.