Dates and verification labels belong to the original edition. A past verified announcement can now be outdated. Historical project names are retained; an SN filter searches the slot across registrations. All keywords must match the story text or subnet label. Ratings describe editorial priority, not investment quality.
REAX plans a 10% miner-budget burn from October 30; inference share falls from 50% to 40%
REAX’s announcement, pictured in Shizzy’s October 26 post and corroborated by the dated operator brief, schedules the inference share of its miner reward budget to fall from 50% to 40% after the October 30 engine slot. The freed 10 percentage points of that miner budget are to be burned; this is not a claim that 10% of all subnet emissions or existing holder balances are destroyed. The team says future inference rewards will follow measured customer demand, while its two competitions keep their full rewards for now. This is a planned emission mechanism, not a completed transaction burn or an already-executed change to the five-slot eTAO vector. Separately, Harnyx announced REAX provider support, and the operator brief reports early GPU-cap activation with validator 1.2.0rc11.post7.1.
Useful adoption and operator detail; routine caps and planned reward changes remain three stars.
Ingot corrects its q2 label to 1.75-bit and publishes format work
Ingot’s repository changes label the q2 round as 1.75-bit rather than 2-bit and add publication work for related GGUF outputs. The precision distinction matters when comparing model size and benchmark claims. Code publication does not independently establish model quality or availability of every output.
A meaningful correction to a technical comparison; no unverified performance uplift asserted.
Last night: SN98 NeverPlayAlone changed its allocation to 33% owner, 7% validators and 60% miners
Carried forward from the October 26 edition so today’s morning report includes the overnight change: A successful finalized transaction at block 9,371,009 on October 26 at 23:01:36 UTC changes the five-slot allocation from 0% burn / 0% recycle / 18% owner / 41% validators / 41% miners to approximately 0% / 0% / 33% / 7% / 60%. This increases both the owner and miner shares while reducing the validator bucket. Root receives its live baseline first, capped by that validator allocation; alpha validators share any remainder. These are allocation weights, not guaranteed individual returns, and historical yield estimates can lag the change. Exact vectors, successful execution and historical registration identity are retained in the evidence.
A finalized reward-allocation change affects subnet participants without establishing a new product or promising individual yield.
SN103 changes from Unknown to Terza Inference; owner coldkey migrated
SN103’s finalized owner-submitted identity changed from unnamed (displayed as Unknown) to Terza Inference, a project describing an AI inference marketplace on Bittensor. The identity update is at block 9,375,503, extrinsic 29. Earlier on October 27, a successful ColdkeySwapped event at block 9,374,413, extrinsic 31, migrated the owner address from 5FHb1T…uPzXz to 5HZJ7Q…PweAN. The reviewed registration block remains unchanged. These are verified identity and owner-key events; they do not establish a sale price, independently tested marketplace availability or paying customer usage.
A verified ownership/identity development, without inventing a sale or carrying old-project returns into a new claim.
Beam publishes Bittensor EVM receipts for transfers and Room deliveries
Beam announced its Bittensor EVM receipt integration on October 26. Its documentation identifies a mainnet contract and lets organizations record transfer or Room-delivery metadata, with file payloads staying off-chain. The important limit: these are reports written by Beam’s relay, not independent proof that content is correct; some Room receipts can include recipients’ signatures. This gives applications a documented way to read completion evidence from a contract, without proving that every proposed payment workflow is already deployed.
A documented mainnet integration with clear trust limits and an application use case.
Minos follows up on its Montreal genetics appearance
In an October 26 post, Minos thanks the American Society of Human Genetics and describes sharing its genomic-AI work in Montreal. Its attached graphic explicitly identifies ASHG 2026. This is a team-reported follow-up to the earlier scheduled appearances, recovered in the seven-day overlap because it fell just before the previous report cutoff. It does not independently establish exact session times, new laboratory customers, or an investment or commercial partnership with OpenAI.
Verified dated team follow-up or finalized technical housekeeping; no unsupported business outcome inferred.
SOMA reports Claude Code support and more model options
SOMA’s October 26 update reports Claude Code support and additional models, plus an increase in its measured token savings from 30% to 40% after its first Jev round. The savings figure is the team’s own claim, with no independent workload comparison performed here. Model availability and compression quality should be checked for the intended task; a token reduction alone does not establish equal answer quality or a universal bill reduction.
A useful product update, with performance claims kept attributed rather than promoted as verified benchmarks.
Ditto reports evaluation changes and a withholding split
Ditto’s October 27 brief reports evaluation wording changes and a 15% generator/35% agent split with 50% withheld from epoch 25,962. That wording concerns the incentive mechanism; this edition does not count it as a qualifying transaction burn. The next integration version is discussed as work in progress.
A useful operator/evaluation change; withholding is explicitly excluded from transaction-burn totals.
Attelier describes a staged video-inference bootstrap
Attelier’s October 27 brief describes a permissioned video-inference bootstrap, a provider-application window and a later gateway step. It says weights are not yet active and trace verification remains work in progress. These are staged rollout plans, not a completed open network or measured customer deployment.
A potentially material rollout, accurately kept developing until the activation evidence exists.
Taoswap announces a Bittensor–Ethereum bridge
Taoswap’s October 26 announcement says users can route Ethereum USDC or ETH into TAO or subnet tokens, and convert TAO back to USDC or ETH, powered by tao.com. It offers another entry route into subnet markets. This is a team-reported product availability announcement: this review did not execute a bridge transfer, audit the bridge’s security, or verify volume and settlement performance.
A material new access route for the ecosystem, qualified to the launch evidence actually reviewed.
Runtime 486 is finalized; the reviewed delta concerns stake availability reads
The daily scan observes finalized runtime 486. The reviewed release source updates weight handling for stake-availability reads; this review does not claim a new allocation or burn rule from the version number alone. The collectors retained runtime-specific event layouts across the window rather than silently decoding every block with a single version.
A verified protocol-version observation with a bounded source interpretation.
ForeverMoney reports 13,393.6 TAO bridged since launch
ForeverMoney’s October 27 update reports cumulative bridging of 13,393.6 TAO in under 80 days: 8,521.78 to Base and 4,871.82 to Robinhood Chain, valued by the team at about $3.32 million in total. This is a cumulative activity milestone, not current bridge TVL, audited revenue or a new bridge launch. The attached animation shows a separate bridge-TVL series; it should not be equated with the cumulative amount bridged. The team describes native 1:1 TAO movement using Chainlink CCIP; this review did not execute a transfer or audit reserves/security.
A relevant team-reported usage milestone with cumulative flows separated from current holdings and security claims.
SN8 Vanta records two qualifying alpha burns totaling 700.188415751α
Two successful finalized AlphaBurned events on October 26 destroyed 533.521749085α at 04:02 UTC and 166.666666666α at 05:04 UTC. Each exceeds the 100-alpha reporting minimum and belongs to SN8 Vanta’s historical registration. Neither has a matching AddStakeBurn marker, so these are direct alpha burns, not independently proven revenue-funded buybacks. They are subnet burns from the executing wallet, not automatically burns by the current owner coldkey.
Exact finalized transactions establish supply destruction without inferring revenue or price effects.
SN12 ComputeHorde publishes dependency hardening for a CA incident
A merged October 26 commit labelled eMudhra CA incident immunization updates internal dependency locks across executor, validator and related components, including bittensor-pylon-client 2.3.1. This establishes published maintenance, not that an SN12 operator was compromised or every deployment has updated. Operators can inspect the pinned versions and normal upgrade guidance.
Security-related maintenance is relevant, but the commit alone does not establish a subnet breach.
SN14 Cacheon burns 2,843.585 alpha in a finalized transaction
A successful AlphaBurned transaction at 2026-10-26T19:59:24.001000+00:00 destroyed exactly 2,843.585 SN14 alpha (2,843,585,000,000 native units) at block 9,370,098, extrinsic 19. Historical on-chain identity confirms Cacheon, registered at block 4,848,444. There is no matching AddStakeBurn event in this extrinsic, so this is a direct alpha burn, not evidence of a revenue-funded buyback. Routine transaction fees are excluded.
Finalized execution verifies meaningful token destruction, without establishing revenue, adoption or future price effects.
SN25 UR and Persistent AI announce a Private Inference integration
UR says Persistent AI, through Private Inference, is joining as a network operator. Private Inference independently says it will become an operator and route traffic through the proxy network to address connection-level privacy. Both sides announced the relationship; the partner’s future-tense wording means traffic migration and a measured usage increase are not marked complete. UR says operators pay for traffic in its alpha, but no purchase volume or token-price effect is established.
Both parties confirm a specific integration plan; actual traffic activation remains to be verified.
SN28 SayGM publishes a Taostats adoption case study
A SayGM case study dated October 25 and promoted October 26 describes Taostats rebuilding doug as a live-chain-data agent with its model layer routed through SayGM. The provider reports more than 33,000 questions, failed model calls falling from 1.29% to 0.085%, and costs about 12% below the compared Anthropic list price. The article includes attributed customer quotes. These are case-study figures, not our independent billing or reliability audit, and do not establish a return to alpha holders.
A named customer and specific reported operating results provide substantive adoption evidence, with measurement limits explicit.
SN34 BitMind prepares a benchmark refresh and model-only rules
The October 26 team-attributed digest says submissions are closing ahead of a benchmark refresh and decision logic must reside in the loaded safetensors model, without extra hard-coded rules. The team also acknowledges evaluation issues and says fixes are coming. This is a forthcoming participation-policy change; exact activation and completed recovery were not independently established.
A useful participation-rule change belongs in short updates, with plans separated from execution.
SN64 Chutes sunsets subscriptions; new models are PayGo-only beta
Chutes says active subscriptions will continue until their current billing period ends, then will not renew. The team cites new model economics and the GPU market. It has begun rolling out Qwen3.8 Max and DeepSeek V4.1 Pro in beta, available only on PayGo. Crucially, these two offerings do not currently support TEE, confidential-compute or end-to-end-encrypted mode. Users relying on subscriptions or those privacy modes need to check their model and billing configuration; this does not mean every Chutes model lost confidential support.
A billing change and explicit privacy limitations affect existing users, alongside a concrete model rollout.
SN65 True Performance Network teases a November 2 announcement
The official October 26 post says Local, Private, Free and gives November 2, 2026 as a date. Its attached video contrasts recurring subscriptions with local-AI messaging. A product name, download, hardware requirements and release time are not established. Treat this as a dated item to watch, not a completed launch.
A concrete date makes the teaser worth tracking without supporting a product-launch claim.
Searches published headlines and summaries, not private bookmarks, unpublished leads or source-page contents. Duplicate headlines within an edition appear once; separate editions remain separate so follow-ups are preserved. Share the page URL to retain your filters.
