IOTA distributes LLM pretraining across thousands of devices including a Mac app.
The simplest explanation
Like SETI@home but for training AI: your laptop earns tokens for helping build the model.
No jargon. No whitepapers. Just the facts.
Training frontier AI models costs hundreds of millions of dollars and is only possible for OpenAI, Google, and a handful of other companies. The rest of the world uses whatever those companies decide to release. There has never been a permissionless way to participate in building foundational AI.
IOTA distributes LLM pretraining across thousands of devices including a Mac app.
IOTA (Incentivised Orchestrated Training Architecture, by Macrocosmos) uses pipeline-parallel training to distribute model layers across miners, streaming activations between them — enabling model sizes that exceed any single GPU's VRAM. A 'Train at Home' initiative lets consumer GPU owners contribute with zero ML knowledge required. Model size scales with the number of participants, not individual VRAM, making it adversarially robust even with untrusted nodes.
The best way to understand a new technology is to compare it to something familiar.
IOTA (Incentivised Orchestrated Training Architecture, by Macrocosmos) uses pipeline-parallel training to distribute model layers across miners, streaming activations between them — enabling model sizes that exceed any single GPU's VRAM. A 'Train at Home' initiative lets consumer GPU owners contribute with zero ML knowledge required. Model size scales with the number of participants, not individual VRAM, making it adversarially robust even with untrusted nodes.
iotauses Bittensor's incentive layer to build something no single company could run alone.
Participants (called miners) on the iota subnet compete to produce the best outputs for ai model training tasks. Anyone with the right hardware can join.
Validators continuously evaluate miner outputs against objective benchmarks. For iota, this includes LLM benchmark (MMLU, HellaSwag). There's no human committee — the protocol decides.
Miners are paid in the iota alpha token in proportion to how good their work is. This creates a continuous competitive pressure that drives quality up and cost down — structurally, not just as a promise.
Because every participant is aligned toward the same goal — producing the best ai model training results — the network improves continuously without requiring a central team to manage it.
iota is Subnet 9 (SN9) on the Bittensor network — a decentralized AI protocol built on the TAO blockchain. IOTA distributes LLM pretraining across thousands of devices including a Mac app.
Training frontier AI models costs hundreds of millions of dollars and is only possible for OpenAI, Google, and a handful of other companies. The rest of the world uses whatever those companies decide to release. There has never been a permissionless way to participate in building foundational AI.
IOTA (Incentivised Orchestrated Training Architecture, by Macrocosmos) uses pipeline-parallel training to distribute model layers across miners, streaming activations between them — enabling model sizes that exceed any single GPU's VRAM. A 'Train at Home' initiative lets consumer GPU owners contribute with zero ML knowledge required. Model size scales with the number of participants, not individual VRAM, making it adversarially robust even with untrusted nodes.
The iota subnet has its own alpha token on Bittensor's dTAO system. It trades in the Bittensor liquidity pool and its price reflects market demand for the subnet's services.
AlphaGap tracks iota using its aGap score — a composite of development activity, token flow, and social signals. This page is for informational purposes only and is not financial advice. Always do your own research before making any investment decisions.
AlphaGap tracks signals, whale flows, developer commits, and the aGap score for every Bittensor subnet — updated continuously. Find the alpha gap before everyone else.
Free preview · Pro from $29/mo · Premium from $49/mo