mikyo.one / IOTA / The IOTA Field Map Updated 29 Sep 2026
02 · The field

The IOTA Field Map

Who else trains AI across machines they don't own, and how each compares with IOTA on mechanism, largest run, and what is sold.

12 projects, IOTA included; 7 have trained a model across compute outside a single reserved cluster. The largest dense model trained across machines on this map is IOTA's Orion-100B. Open a project for its words, its look and its facts.

Covenant AIPrime IntellectGradientsChutesLiumTargonNous ResearchPluralis ResearchFlower LabsGensynTeutonicIOTA
Dot size: largest shared training run outside one reserved cluster. Hollow: none.
For comparisonIOTAMacrocosmos · Bittensor subnet 9100Blargest run outside one reserved clusterRival for the buyerCovenant AITemplar, Basilica, GRAIL · left Bittensor April 202672Blargest run outside one reserved clusterRival for the buyerPrime IntellectHosted RL · GPU marketplace10Blargest run outside one reserved clusterlargest model 106B, 12B activeRival for the buyerGradientsRayon Labs · fine-tuningFine-tuning jobs by UI and API, pay per useRival for the supplyChutesRayon Labs · inference, training research7.8Blargest run outside one reserved clusterRival for the supplyLiumDatura · GPU rentalHourly pods, asks on 29 September: H200 $3ComplementTargonManifold Labs · confidential computeConfidential VMs and sandboxes by consumption: H200 $3ReferenceNous ResearchPsyche (NousNet) · DisTrO · DeMo40Blargest run outside one reserved clusterlargest model 405BReferencePluralis ResearchProtocol Learning · Agora8.6Blargest run outside one reserved clusterReferenceFlower LabsPhoton · SuperGrid · Endeavor7Blargest run outside one reserved clusterReferenceGensynVerde · open-1b · DelphiNo training run can be boughtReferenceTeutonicking of the hill · the old Templar slotNothing

About the field

The frame, the relations, what the words say across the field, and how the page was made.

The frameunder 150 words

A model builder chooses between three things. A reserved cluster: matched hardware, held for months, the highest price. Spot: whatever is free, taken back without notice, the lowest price. Or a swarm: heterogeneous compute the builder does not own or reserve, joined and left mid-run, with the run surviving the machines. Every project on this page sells or studies the third.

IOTA's mechanism, in one sentence: a model is cut into pipeline slices, each slice runs on a GPU Macrocosmos does not own, validators score each miner's work, and the chain pays miners in alpha by score share over a rolling day; the subnet has been on mainnet since June 2025 (relaunched August 2025). The difference from the field is the unit of payment (scored work on one slice, not a whole checkpoint or a trusted node) and the shape of the parallelism (pipeline across single cards, not data-parallel copies of a whole model).

The four relations
Rival for the buyerSells today, or on a stated date, a runtime, a managed run or capacity that a model builder without a reserved cluster would buy instead of IOTA's SDK or managed route. Pretraining, fine-tuning and RL all count; IOTA pitches all three. Raw GPU hours count only weakly.
Rival for the supplyRecruits the same class of card IOTA runs on (Apple Silicon Macs, 24 to 32 GB consumer GPUs, single A100, L40S or A6000 cards), and the owner must choose: the card earns from one or the other. Sharing TAO emission is not supply rivalry.
ComplementHas a concrete path to buy from IOTA, sell compute to it, or distribute it. The card names the path. None of these paths is a public arrangement unless the card says so.
ReferenceLineage, a benchmark or a competing claim: IOTA inherits from it, is measured against it, or is compared with it. It has no offer a builder could buy instead of IOTA and draws on no shared supply today.
IOTA, in three paragraphs

Mechanism. Pipeline-parallel across single GPUs, 16 stages and 3 replicas at 100B. An orchestrator assigns slices; validators score each miner's forward-pass work, and miners are paid in alpha by score share over a rolling window of about a day, with a cap and the excess burned. Control is central, the data path is peer to peer (Iroh since v3.0.0, 9 March 2026). The open way to join today is Train at Home on an Apple Silicon Mac with 16 GB; the GPU miner repository still lists 16 GB of VRAM. Runs survive nodes leaving; the second 16B run is the repeatability claim.

Largest run. Orion-100B: 100B parameters, 1.1B tokens, 48 non-colocated A100 80GB across five US datacenters, "provisioned from multiple providers" (unnamed), about two days, 30.8% average MFU, 65% of co-located speed, $20 a replica-hour; a viability run, stopped for cost. Orion-16B: 256 GPUs at launch (27 July), c.180+ (5 August), 225 across three continents (20 August), then a mixed roster of 4090s, 5090s, A100s, L40S and A6000s, finished 23 September; GPUs from "independently operated suppliers", one named (Green Compute); 100B tokens by 14 August, about 20% MFU. How Orion suppliers were paid is not published. No checkpoint offered for download.

Sold, and to whom. The iota SDK and Liquid Compute, announced 28 September 2026: early access by registration, the SDK page still "coming soon". Will Squires says the SDK covers pretraining, RL and fine-tuning; published results are on the legacy code base. Today the door is a managed route ("You own the outcome. We run the training"), by contact form, no pricing. The site also names GPU operators with idle capacity as SDK users. No named paying customer as of this page. Funded by emission and OTC alpha sales; Stillcore Capital invested in SN9 in May 2026 (secondhand, terms unknown). 164 companies in conversation was said from the Exploit stage and is secondhand.

Every number above is on the register with its label and source.

What the words say across the field

Common words. "Open" is on nine of twelve home pages, including IOTA's. "Decentralized" is on six. "Distributed" on four.

Verification words. No home page on the map uses "verifiable", "auditable" or "trustless". "Incentivized" appears only on Templar's, which is dormant. IOTA's own name already contains it.

Numbers on home pages. Three rivals lead with a number on the home page: Pluralis ("6B-parameter open pretraining run trained on 500B tokens across 330 contributors", their home page), Prime Intellect ("INTELLECT-3: A 100B+ MoE"), Chutes ("Trillions of Tokens per Month"). IOTA's page leads with none; Orion-100B, the largest dense model trained across machines on this map, is not on its own home page.

How the founder-voice reads were made

Founder-led marketing means the founder's own account and appearances carry the message and the brand account amplifies them. This section records, for IOTA and for every entity on the map, who actually carries it: the canonical accounts with their size and cadence, the founders' bios and dated statements across X, podcasts, talks, essays and interviews, and the words each person keeps using. Quotes are verbatim, under twenty words, read from the source on 28 September 2026; X timelines were read through mirrors or the browser, so thirty-day cadence is marked where it could not be observed.

The pattern across the field. Three shapes. Founder-led: Steeves (30,759 followers, no brand channel at all), Long (the argument on his account, metrics on the company's), Dare before April. Brand-led: Prime Intellect (84,789 on the company account, every post signed "Team"), Gensyn (93,020, retweeting its founders), Nous on launches (275,027), Targon and Flower (no founder byline in 2026). Private founder-led: Fish, who sets Lium's prices by Discord directive and has never appeared on a podcast. The one founder whose reach matches the brand is Teknium (130,042 against 275,027), and he is the product's hands, not its CEO.

The words. Steffen's adjectives: "heterogeneous, unreliable, permissionless and token incentivized" (May 2025); "globally distributed, heterogeneous and unreliable" (September 2026). "Trustless" appears in 2025 posts and the docs and in nothing from 2026. "Liquid" replaced "swarm" in June; Pluralis's founding scientist uses "liquid compute" too (May and September). "Compute nobody reserved" is in the IOTA account's own copy.

Reference striplineage, not rivals
SWARM Parallelism, 2023
Ryabinin, Dettmers, Diskin, BorzunovTemporary randomized pipelines over preemptible T4s at 200 Mb/s, rebalanced on failure, trained a 1B shared-parameter model. IOTA's nearest ancestor. IOTA adds a central orchestrator that assigns, scores and pays on chain. source
Petals and Hivemind, 2020 to 2022
learning-at-home; Borzunov et al.Volunteer machines found over a DHT; Petals served and fine-tuned BLOOM-176B by splitting layer blocks across consumer GPUs. The same shape as IOTA's slices, with trusted peers and nobody paid. source
DiLoCo, 2023; Streaming DiLoCo, 2025
Douillard et al., Google DeepMindData-parallel replicas take many local steps and sync through an outer optimizer, matching synchronous training with 500x less communication; streaming subsets cut bandwidth two more orders. The method behind Prime, Nous, Covenant and Flower. IOTA shares the slow-links premise, not the data-parallel shape. source
DeMo, 2024
Peng, Chen, Su, Quesnelle, Kingma, Liu (Nous)Decoupled momentum with a sparsified transform sends up to 85x less data per GPU than AdamW-DDP at 300M to 1B. Compresses gradients between replicas; IOTA compresses activations between slices. Different bottleneck, could stack. source
Decentralized training in heterogeneous environments, 2022
Yuan et al.; Together's foundersScheduling for geo-distributed runs across three continents, 4.8x over Megatron. The academic root of the whole field; the company now sells clusters. source
Considered and left out
Affine, SN120A teacher-anchored distillation contest, not distributed training: each miner fine-tunes its own model to match a frozen teacher; nothing sold. The operator rents Lium pods to serve the leading model. Steeves matters as a person and is on the Teutonic card.
Swarm, SN124Drone flight policies, one per miner, trained locally and scored in simulated worlds; no LLM, no compute sold, no fleet. A €370k raise and a solar-park pilot, secondhand.
Bagel LabsNow a physical-AI lab. Paris (2025) trained 8 isolated experts with no communication on clouds it rented (AWS, GCP, Runpod, local); no network, nothing sold; Bakery is offline. Toronto; $3.1M pre-seed (2024), secondhand.
EXO LabsLocal inference on devices you own, sold to businesses as deployments; no training product. Training work is small-cluster and simulated research (SPARTA, 124M, up to 8 nodes; EXO Gym idle since December 2025). No open network.
io.net, Akash, Hyperbolic, Aethir, Render, TogetherNone trains one model across machines it does not run as one cluster. Akash, Hyperbolic, Aethir and Render rent GPUs and the customer runs the job. io.net (Training as a Service, beta) and Together sell managed fine-tuning on a rented cluster, not pretraining. They rent out other people's hardware.
ChronoLLM SN38; SN39, Basilica's former subnetSN38 is ChronoLLM: miners train whole models with a date cutoff and a validator duels them; not distributed training, nothing sold. Emission 3.9% today. SN39 has been deprecated since Covenant's April exit and carries no emission; Basilica now runs off-chain and is on the Covenant card.
The claim most likely to be disputed
Most likely to be raised by either teamWhich run is the largest. By tokens, Covenant-72B: about 1.1 trillion against Orion-100B's 1.1 billion. By dense parameters, Orion-100B (Teutonic-II has 110B total but is a mixture-of-experts with about 7.3B active, trained one checkpoint at a time). Covenant-72B was a completed permissionless pretraining run; Orion-100B was a viability run on 48 provisioned A100s, stopped for cost; the permissionless run is Orion-16B. The Covenant-72B paper (arXiv 2603.08163, 1.09T DCLM tokens plus annealing) beside the Orion-100B write-up (1.1B tokens, stopped for cost). Both numbers are on the page; the page says which is which.
How an entity is added

An entity is added when it meets the inclusion rule and a primary source (its own docs, paper, repository, or announcement) can be opened for its mechanism. Every card carries the same eight fields. Every cell is filled or marked unknown. A company is never called a customer, partner, or pilot of IOTA unless Macrocosmos has said so in public.

Mechanism and runs are taken from primary sources only. Funding and headcount may come from aggregators or press, labeled.

The page is rendered from field.json by tools/render_field.py. Fork the data. Send corrections to connect@mikyo.one and they go in with a date.

Assumptions on this page
  • An entity is in if it trains one model across machines that are not run as one cluster and exchange updates during the run, sells such a run or a runtime for one, or is a Bittensor subnet that pretrains a language model from scratch or sells compute or training. Pure inference marketplaces are out unless they have announced such training.
  • Relation is judged against the buyer IOTA names (a model builder without a reserved cluster, for pretraining, fine-tuning or RL) and the cards IOTA runs on (Train at Home Macs, consumer GPUs, single datacenter cards), as of 29 September 2026. The tests are in the four relations above.
  • Funding and headcount come from the 6 September 2026 landscape scan (`landscape/decentralized-ai.csv`) unless a card gives a newer source. They are aggregator or press figures and are labeled so.