Skip to main content

IOTA (Macrocosmos AI)

IOTA (Macrocosmos AI) — a reference card for a crypto project: what the product does and how verifiable the project’s key claims are; also — alignment with the Decentralized AI thesis (DecAI Fit).

IOTA (Macrocosmos AI) — an AI/Training \| Infra project on the Bittensor network (Subnet 9).

Disclaimer

Important: this page is a reference overview of the project. Within DecAIHub, we separately record a verification based on open sources within the AI thesis (what is real AI/onchain/why a token, and which sources this is based on).
This is not financial advice and not a recommendation to buy/sell/hold any asset. We do not assess price, returns, or entry/exit timing.
Conclusions are based on publicly available materials found as of the verification date. The absence of confirmation does not imply the absence of the technology.

TL;DR (for readers who are not into AI)

  • What it is: a subsystem/network for decentralized pretraining of large language models (LLM) on Bittensor.
  • Who it’s for: the subnet’s miners/validators and developers who want to participate in distributed training.
  • Why onchain/token: onchain is used for incentives/accounting of participants’ contributions (the logic of Bittensor subnets).
  • What you can actually touch (Tier‑1): public code + a tech primer; a network state dashboard.

How it works (very briefly)

  1. The participant runs a miner: takes on a part of the training computations (layers/activations).
  2. The orchestrator distributes the work: hands out layers/batches and manages synchronizations.
  3. Validators verify the contribution: spot‑check/reproducibility of computations and accounting of the volume of work.
  4. Where the blockchain is: the Bittensor layer — for incentives/accounting and emission distribution among participants.
  5. Token role: incentivizing participants’ work within the Bittensor subnet mechanics.

Quick links: AI thesis fit

Project passport

Field Value
Project name IOTA (Macrocosmos AI)
Also known as (aliases) Subnet 9 IOTA, SN9; Incentivized Orchestrated Training Architecture
Logo (logo) ../assets/458_iota-macrocosmos-ai_logo.jpg
Token symbol (ticker) IOTA (Subnet 9 alpha token)
Category AI/Training | Infra
Primary chain (chain) Bittensor
Type Appchain
Network status mainnet
Consensus (consensus) —
Launch date (launch_date) —
Project status (status) active
Token (standard) native
Token contract Subnet 9 (SN9)
Token model (token_type) utility (incentives)
Max supply —
Issuance —
Tokenomics (tokenomics, brief) Rewards/emissions are distributed according to miners’ contribution; validation via spot‑check and reproducibility.
Governance —
Admins/keys —
Audit —
Bug bounty —
Team (team) Macrocosmos AI
Funding (investors/funding) —
AI component (ai_component) distributed LLM pretraining (IOTA)
Use‑case (use_case) decentralized LLM pretraining on Bittensor
DecAI Fit — AI thesis fit 5/6 — high alignment with the Decentralized AI thesis
DecAI Fit component summary ai_score 3 / onchain_score 2 / token_score 1 / evidence_score 3
Last verified 2026-02-11

General project overview

IOTA (Incentivized Orchestrated Training Architecture) is an architecture/framework for distributed LLM pretraining designed for a network of “heterogeneous and unreliable” participants, where the contribution is measured and rewarded via the subnet mechanics. In the Tier‑1 tech primer, the goal is formulated as scaling training through a data‑ and pipeline‑parallel approach with an orchestrator, validation and synchronization mechanisms.

Features/modules

  • Orchestrator: distribution of layers/work and management of synchronizations.
  • Miners: perform computations on “their own” section of the pipeline.
  • Validators: reproducibility/spot‑check, accounting of contribution and anti‑exploit logic.

Typical scenarios

  • Scenario 1: a participant runs a miner/validator → participates in training and receives rewards based on contribution.
  • Scenario 2: a researcher reads the tech primer → understands the IOTA architecture and its limitations (communications/validation/synchronization).
  • Scenario 3: a developer makes changes to the code → improves parts of the pipeline/validation.

Getting started (wallet/network)

  • What you need: an environment for running a miner/validator (software from the repository) and access to the subnet infrastructure.
  • Network/chain: Bittensor (Subnet 9).
  • Fees: network costs/participation conditions depend on the subnet mechanics and the environment (see Tier‑1 sources).

Architecture/tech (non‑AI)

  • Components: orchestrator, miners, validators, shared storage (for example S3 for weights/activations), communications between pipeline stages.
  • Participant roles: miners, validators, supporting services/storage.
  • Where logic runs: training/validation — offchain at the participants; incentives/accounting/emissions — in the Bittensor onchain layer.
  • Dependencies: storage and network channels for activations/weights; a runtime environment (GPU/software).

Token/economics (expanded, no forecasts)

  • Token utility: participants receive rewards for computations/validated contribution to training.
  • Protocol necessity drivers: incentivization is needed for permissionless participation and accounting of contribution at network scale.
  • Issuance/burn: not established per Tier‑1 (no numerical parameters of alpha token issuance/supply were found in the subnet’s public documentation).
  • Security/incentive mechanisms: validators reproduce a portion of the computations, and rewards are tied to measurable contribution.
  • Limitation: the supply/issuance parameters of the alpha token and other numerical tokenomics parameters are not disclosed in the available Tier‑1 materials of the subnet.

Governance

  • Who makes decisions: not established per Tier‑1.
  • How protocol parameters change: —.
  • Where decisions are discussed/made: GitHub (issues/PR) and public updates (see links_*).
  • Keys/multisig: —.

Security and incidents

  • Audits: not established per Tier‑1.
  • Bug bounty: not established per Tier‑1.
  • Incidents: not established per Tier‑1.
  • Threat model: malicious/dishonest participants are possible; the Tier‑1 primer highlights verification/exploit detection mechanisms in the context of a trustless environment.

Project status and timeline

  • Current status: active (see the public code/updates).
  • Key releases/milestones: not established per Tier‑1 (there are public updates/posts and changelog surfaces in links_*).
  • Roadmap: the pre‑training tag/posts on Substack (see links_*).

Ecosystem/integrations

  • Integrations: —.
  • Partners: —.
  • Where it’s actually used: participation in the Bittensor subnet (Subnet 9) and running miners/validators; state visualization — via the dashboard (see links_*).

AI component of the project

  • What is real AI here? the Tier‑1 primer describes distributed LLM training (pretraining) and components/algorithms designed for a trustless and heterogeneous environment.
  • What is truly onchain? incentives/accounting of participants and emissions within the Bittensor subnet.
  • Why is a token needed? incentivizing miners/validators so that the network can sustainably train models in a permissionless mode.
  • Red flags (based on the information found): “frontier‑scale” promises without reproducible artifacts; unclear tokenomics/reward parameters without Tier‑1 specifications; absence of public code/a primer (here they exist).

DecAI Fit component summary

  • DecAI Fit: 5/6 — high alignment with the Decentralized AI thesis
Component Level Score
AI reality high 3
Onchain necessity medium 2
Token necessity low 1
Evidence Tier‑1 present 3
Details

AI reality

AI_0 — AI intent (Tier‑1 roadmap) (1/0)

  • AI_0 (decision): 0
  • Claim (one line): Tier‑1 sources describe a working AI component (LLM training), not just “intent”.
  • Evidence: Tier‑1 — https://iota.macrocosmos.ai/research/iota_primer.pdf
  • Why sufficient / why not (one line): AI_0 applies when there is only intent; here there is a tech description of the training.
  • Conflict note (if any): —

AI_1 — what the AI does (1/0)

  • AI_1 (decision): 1
  • Claim (one line): The system is intended for distributed LLM pretraining via a data‑ and pipeline‑parallel architecture.
  • Evidence: Tier‑1 — https://iota.macrocosmos.ai/research/iota_primer.pdf ; https://github.com/macrocosm-os/iota
  • Why sufficient / why not (one line): The Tier‑1 primer and the repository directly describe the task and the training architecture.
  • Conflict note (if any): —

AI_2 — demos/examples (1/0)

  • AI_2 (decision): 1
  • Claim (one line): There is a public dashboard showing the state of the network/participants/subnet metrics.
  • Evidence: Tier‑1 — https://iota.macrocosmos.ai/dashboard/mainnet
  • Why sufficient / why not (one line): The dashboard is a reproducible “surface” for observing the network’s operation and participation.
  • Conflict note (if any): —

AI_3 — metrics/quality (1/0)

  • AI_3 (decision): 0
  • Metrics: —
  • Context (version/date/how measured): —
  • Evidence: Tier‑1 — https://iota.macrocosmos.ai/research/iota_primer.pdf
  • Why sufficient / why not (one line): In the Tier‑1 materials, no unified set of model quality metrics (benchmarks/perplexity etc.) with context (version/date/dataset) was found, hence conservatively 0.
  • Conflict note (if any): —

AI_4 — tech description/architecture/versions (1/0)

  • AI_4 (decision): 1
  • Claim (one line): The IOTA tech primer has been published with the architecture (orchestrator/miners/validators), synchronizations and mechanisms (compression/merge etc.).
  • Evidence: Tier‑1 — https://iota.macrocosmos.ai/research/iota_primer.pdf
  • Why sufficient / why not (one line): The document contains a structured tech description and roles, sufficient for AI_4.
  • Conflict note (if any): —

Onchain necessity

OC_1 — contracts/addresses/chain (1/0)

  • OC_1 (decision): 1
  • Chain: Bittensor (Subnet 9)
  • Addresses/identifiers: Subnet 9 (SN9) on Bittensor
  • Evidence: Tier‑1 — https://iota.macrocosmos.ai/research/iota_primer.pdf
  • Why sufficient / why not (one line): The Tier‑1 primer directly records operation in the context of Bittensor Subnet 9.
  • Conflict note (if any): —

OC_2 — onchain beyond just a token (1/0)

  • OC_2 (decision): 1
  • What useful logic is onchain: accounting/distribution of incentives and rewards to subnet participants.
  • Evidence: Tier‑1 — https://iota.macrocosmos.ai/research/iota_primer.pdf
  • Why sufficient / why not (one line): The Tier‑1 source describes a “token incentivized” environment and accounting of contribution via the subnet.
  • Conflict note (if any): —

OC_3 — onchain is critical (1/0)

  • OC_3 (decision): 0
  • “Replace the blockchain” test: no Tier‑1 justification was found that without the blockchain the system “breaks down” (rather than becoming a centralized coordination service).
  • Evidence: Tier‑1 — https://iota.macrocosmos.ai/research/iota_primer.pdf
  • Why sufficient / why not (one line): OC_3 requires an explicit argument “a key property is lost without onchain”; in the available Tier‑1 sources for the subnet this is not formulated.
  • Conflict note (if any): —

Token necessity

TK_0 — no token (1/0)

  • TK_0 (decision): 0
  • Claim (one line): SN9 participants receive rewards in IOTA alpha tokens (the subnet/alpha token of Subnet 9).
  • Evidence: Tier‑1 — https://docs.macrocosmos.ai/subnets/subnet-9-iota/mining
  • Why sufficient / why not (one line): The Tier‑1 subnet documentation explicitly describes rewards in the alpha token, hence the token exists.
  • Conflict note (if any): —

TK_1..TK_4 (1/0)

TK_1 — payment for the service (1/0)

  • TK_1 (decision): 0
  • Claim (one line): No Tier‑1 description of “payment for the service” with the token as a mandatory mechanism for users was found.
  • Evidence: Tier‑1 — https://iota.macrocosmos.ai/research/iota_primer.pdf
  • Why sufficient / why not (one line): In the available Tier‑1 sources of the subnet, no explicit pay‑per‑use model was found.
  • Conflict note (if any): —

TK_2 — incentives/collateral/penalties (1/0)

  • TK_2 (decision): 1
  • Claim (one line): Miners receive rewards (incentives) in IOTA alpha tokens; the documentation describes the incentive mechanism and reward structure.
  • Evidence: Tier‑1 — https://docs.macrocosmos.ai/subnets/subnet-9-iota/mining ; https://docs.macrocosmos.ai/subnets/subnet-9-iota/subnet-9-incentive-mechanism
  • Why sufficient / why not (one line): Tier‑1 describes the incentives and the reward mechanism for miners; this is sufficient to consider the token used for incentives.
  • Conflict note (if any): —

TK_3 — governance (1/0)

  • TK_3 (decision): 0
  • Claim (one line): Governance use of the token is not established per Tier‑1.
  • Evidence: Tier‑1 — https://iota.macrocosmos.ai/research/iota_primer.pdf
  • Why sufficient / why not (one line): In Tier‑1 sources there is no explicit voting/governance via the token.
  • Conflict note (if any): —

TK_4 — token is replaceable (1/0)

  • TK_4 (decision): 0
  • Claim (one line): The replaceability of the token/incentives with another mechanism is not established per Tier‑1.
  • Evidence: Tier‑1 — https://iota.macrocosmos.ai/research/iota_primer.pdf
  • Why sufficient / why not (one line): TK_4 requires an explicit argument; in the available Tier‑1 sources of the subnet it is not found.
  • Conflict note (if any): —

Evidence (table)

Tier Source What it proves
Tier‑1 https://iota.macrocosmos.ai/research/iota_primer.pdf Tech description of IOTA and the AI component (LLM training) + operation in a trustless/permissionless environment with token incentives.
Tier‑1 https://docs.macrocosmos.ai/subnets/subnet-9-iota Overview of SN9 IOTA (purpose, roles, links to the primer and the dashboard).
Tier‑1 https://docs.macrocosmos.ai/subnets/subnet-9-iota/subnet-9-incentive-mechanism Tier‑1 description of the incentive mechanism/reward structure (fixed compensation per activation).
Tier‑1 https://github.com/macrocosm-os/iota Public code (miner/validator/scripts), confirmation of the implementation.
Tier‑1 https://iota.macrocosmos.ai/dashboard/mainnet Network/participants dashboard (what can be observed/verified).
Tier‑2 https://taostats.io/subnets/9/chart Subnet metrics/visualization (context, navigation).
Tier‑1 https://docs.macrocosmos.ai/subnets/subnet-9-iota/mining Miners’ rewards in IOTA alpha tokens (confirmation of the existence of the subnet token and incentives).

No comments to display

Back to top