Macrocosmos
Macrocosmos operates decentralized AI infrastructure on the Bittensor blockchain, running three subnets: IOTA for distributed LLM pretraining, Apex for algorithmic competitions, and Data Universe (Gravity) for large-scale social media data collection. It serves AI/ML teams, enterprises with optimization challenges, and individual compute contributors globally.
- Company typePrivate
- Founded2024
- HeadquartersLondon, United Kingdom
- Headcount11–50
- GTM typeB2B and B2C
- OfferingSoftware
What Macrocosmos does
Macrocosmos Ltd. is a privately held, London-based (22 Chancery Lane, WC2 1LA) AI infrastructure company incorporated in 2024, operating a decentralized AI stack built on the Bittensor blockchain protocol. The company operates three subnets: Subnet 1 (Apex), a competition/routing layer where a global network of miners and agentic agents compete to solve optimization problems such as compression, forecasting, and systems design, with deterministic sandboxed evaluation and automated reward distribution; Subnet 9 (IOTA), a data- and pipeline-parallel training architecture that enables pre-training of frontier-scale LLMs (700M–14B+ parameters) across heterogeneous, unreliable, trustless devices; and Subnet 13 (Data Universe, fronted by Gravity), a real-time social media data collection platform scraping X, Reddit, and YouTube, currently holding 55+ billion rows with up to 80 million new rows added daily. Supporting products include the Training at Home (TAH) consumer desktop app for individual compute contributors, the Nebula 3D visualization engine, the Gravity Marketplace for curated datasets, and the Apex CLI developer tool.
The company's technology portfolio is differentiated by the IOTA training architecture (validated August 2024), the Apex game-theoretic competition framework, and the Gravity data flywheel, with integrations into X, Reddit, YouTube, and the OpenClaw/ClawHub skill ecosystem. Macrocosmos serves three primary customer segments: AI/ML teams and researchers needing fresh, large-scale social data for training and market intelligence; enterprises with quantifiable optimization challenges seeking crowdsourced solutions via Apex; and individual contributors who monetize spare GPU/CPU via the TAH app. Secondary segments include marketing/brand teams for sentiment tracking and research labs seeking crowdsourced benchmark progress.
Monetization combines multiple mechanisms: Gravity subscription tiers (Basic free with $5 credit, Astronaut $99/month, Cosmonaut $499/month, Enterprise custom), pay-as-you-go data pricing ($0.05–$0.10 per 1K records), Gravity Marketplace transaction commissions, Apex competition fees or revenue-share with winning solutions, and Bittensor TAO token incentives earned for subnet operation. Go-to-market is primarily product-led growth through self-serve signup at app.macrocosmos.ai with API-first access, supplemented by sales-led enterprise outreach for custom Apex competitions. Macrocosmos is GDPR-, CCPA-, and UK Data Protection Act-compliant for its Gravity product and markets globally with no geographic restrictions.
Macrocosmos firmographics
Firmographics- Name
- Macrocosmos
- Legal name
- Macrocosmos Ltd.
- Website
- https://macrocosmos.ai
- Company type
- Private
- Founded year
- 2024
- Operating status
- Operating
- Headcount range
- 11–50 employees
- Short description
- Macrocosmos operates decentralized AI infrastructure on the Bittensor blockchain, running three subnets: IOTA for distributed LLM pretraining, Apex for algorithmic competitions, and Data Universe (Gravity) for large-scale social media data collection. It serves AI/ML teams, enterprises with optimization challenges, and individual compute contributors globally.
- Ownership category
- akta.pro rank
Macrocosmos industry classification
Industry- Product category
- Decentralized AI Infrastructure
- NAICS
- Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518210), Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518)
- SIC
- Services-Computer Programming, Data Processing, Etc. (7370), Services-Computer Programming Services (7371), Security & Commodity Brokers, Dealers, Exchanges & Services (6200)
- akta.pro primary industry
- Decentralized GPU/AI Compute Networks (training/inference, GPU renting, model hosting) (FSAPALAC)
- akta.pro secondary industries
- DePIN Orchestration, Scheduling & Middleware (resource discovery, job scheduling, workload orchestration) (FSAPALAK), DePIN Payments, Incentives & Reputation (token incentives, staking/slashing, QoS/reputation, settlement) (FSAPALAL), Network Analytics, AIOps & Root-Cause Correlation (HDAFAGAM)
Keywords
Where Macrocosmos is headquartered
LocationHeadquarters
- HQ city
- London
- HQ country
- United Kingdom
- HQ region
- Europe
Offices1 record
Markets served
Macrocosmos business model
Business model- GTM type
- B2B and B2C
- Offering type
- Software
- Cost components
- Personnel, Technology or R&D, Infrastructure, Operations, Marketing or Sales, Others
Revenue model
- Gravity Data Subscription Plans: Monthly subscription tiers (Basic, Astronaut, Cosmonaut) providing monthly data credits and discounted per-record rates. Higher tiers include more monthly credits and lower per-record pricing.
- Gravity Pay-as-you-go: Pay-per-record pricing for users who prefer not to commit to monthly subscriptions. Records priced at $0.05/1K in-app and $0.10/1K via API for non-subscribers.
- Gravity Marketplace: Platform for purchasing ready-made datasets curated by others. Users browse, buy, and integrate datasets directly into their workspace.
- Apex Competition Fees: Organizations define optimization problems; network of miners competes to solve them. Revenue derived from competition structuring fees or revenue share arrangements with winning solutions delivered to customers.
- Bittensor Network Incentives: Operating subnets on Bittensor generates token incentives (TAO) for network participation, validation, and data contribution activities.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Usage-based | Pay-as-you-go | Basic - Pay as you go for personal use |
| Subscription | Monthly | Astronaut - Small business tier at $99/month |
| Subscription | Monthly | Cosmonaut - Large business tier at $499/month |
| Subscription | Monthly | Enterprise - Custom pricing with advanced features |
Go-to-market motion3 records
Distribution channels6 records
Marketing channels7 records
Macrocosmos product offering
Product offeringCore offering
Macrocosmos builds and operates three decentralized AI subnets on the Bittensor network: Apex (SN1) runs incentive-driven competitions where miners solve algorithmic optimization problems; IOTA (SN9) provides a distributed pre-training architecture that runs across heterogeneous, unreliable hardware; and Data Universe / Gravity (SN13) collects and distributes real-time social media data at scale (55+ billion rows from X, Reddit, and YouTube). Customers access the platform via web apps, Python APIs, a desktop contribution app, and a curated data marketplace.
Product overview
Macrocosmos operates a decentralized AI ecosystem built on three Bittensor subnets: Subnet 1 (Apex) for incentivized algorithmic competition, Subnet 9 (IOTA) for distributed model pre-training, and Subnet 13 (Data Universe) for large-scale social data collection. The core product portfolio includes Apex (an open competition platform where miners submit solutions to optimization problems), IOTA (a distributed training architecture enabling frontier-scale model pre-training across heterogeneous hardware), and Data Universe (a social media data platform accessible via the Gravity scraper, API, and Marketplace). Supporting products include Training at Home (a consumer-facing app for contributing personal hardware to IOTA), Nebula (3D data visualization), and the Gravity Marketplace for curated datasets. Data flows across the ecosystem to power AI model training, real-time social intelligence, and decentralized innovation.
Differentiator
Problem solved
Functional benefit
Brands
- Apex (Subnet 1): A routing layer for intelligence - a competition platform where organizations define problems and a global network of miners competes to solve them.
- IOTA (Subnet 9)
- Gravity
- Training at Home (TAH)
- Nebula
- Data Universe
Products and services
- Data Universe Decentralized data collection and distribution platform built on Bittensor Subnet 13 (SN13), providing access to 55+ billion rows of structured social media data from X (Twitter), Reddit, and YouTube, with up to 80 million new rows added daily. Used for real-time social listening, brand monitoring, market intelligence, and AI model training.
- Gravity Real-time decentralized data scraper that collects social media content from X, Reddit, and YouTube based on user-specified keywords, hashtags, and labels. Accessible via the web UI, the GravityClient Python API, and the Gravity Marketplace for purchasing curated datasets. Targets developers, AI teams, and market intelligence users who need structured social data.
- IOTA (Incentivised Orchestrated Training Architecture) Distributed training architecture and subnet on Bittensor (Subnet 9) designed for data- and pipeline-parallel pre-training of large language models across heterogeneous, unreliable devices in adversarial, trustless environments. Transforms isolated miner competitors into a single cooperating unit that scales arbitrarily while fairly rewarding each contributor. Targets frontier-scale model pre-training workloads (700M to 14B+ parameters).
- Apex Routing layer for intelligence built on Bittensor Subnet 1. Converts clearly-specified optimization problems (compression, control, game, systems-design) into open, incentive-powered competitions where a global network of independent miners and agentic agents submit solutions. Submissions are evaluated in sandboxed environments with deterministic, reproducible scores and automated winner-takes-all reward distribution. Targets enterprises with quantifiable algorithmic challenges.
- Training at Home (TAH) Desktop application that lets individuals contribute their computer's GPU and CPU to the IOTA decentralized AI training network in exchange for token rewards. Available for macOS (Apple Silicon) with Linux support planned; Windows not on the roadmap. Targets individual hardware owners who want to participate in decentralized AI training.
- Gravity Marketplace Marketplace within Data Universe where users can discover and purchase ready-made datasets curated by others. Provides instant access to high-quality structured social media data without needing to configure and run their own scraping tasks. Targets buyers who want pre-curated datasets for immediate integration.
- Apex CLI Command-line interface for miners to interact with the Apex subnet. Allows linking wallets, viewing the competition dashboard, viewing past and current competition submissions, and submitting competition solutions (limited to 4 submissions per day per hotkey). Targets developers running Apex miners on Bittensor Subnet 1.
Quantifiable outcome
- 55+ billion rows of structured social media data available
- +2 more outcomes
Companies that use Macrocosmos
Customer profileNamed customers3 records
Segments5 records
Ideal customer profiles3 records
Macrocosmos technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration4 records
AI capability6 records
Feature5 records
Macrocosmos partnerships and signals
Strategic signalPartnerships
Three partnerships are on record, tiered core and minor.
- ScorecoreScore partnered with SN13's Data Universe to power weather data innovation. This partnership demonstrates how external subnets leverage Macrocosmos' Gravity data infrastructure for their own AI applications.
- Subnet 111coreSubnet 111 leverages Macrocosmos' Gravity for authentic data sourcing. Shows ecosystem collaboration within Bittensor network where other subnets build on Macrocosmos infrastructure.
- TAOLORminorBuilding subnet-native AI agent with distributed RAG using Macrocosmos data infrastructure. Case study demonstrating integration capabilities.
Scale indicators5 records
Recent moves6 records
Expansion highlights6 records
Macrocosmos competitors and assessment
Company assessmentDirect peers
- Render Network: Render Network provides decentralized GPU compute for 3D rendering and increasingly AI workloads on a token-incentivized network. It is directly comparable to Macrocosmos's IOTA subnet as a DePIN-style distributed compute marketplace targeting AI and graphics workloads.
- Gensyn: Gensyn is a decentralized protocol for ML training that coordinates trustless compute across heterogeneous hardware with cryptographic verification. It is a direct peer to Macrocosmos's IOTA on the distributed AI training thesis.
- Akash Network: Akash is an open-source decentralized cloud marketplace for compute, including GPU resources used for AI workloads. It is comparable to Macrocosmos as a permissionless compute marketplace, though with broader workload coverage.
- io.net: io.net aggregates underutilized GPUs globally into a distributed cluster marketed for AI/ML training and inference. It directly competes with IOTA on decentralized AI training compute with similar miner-incentive mechanics.
Emerging players
- Ritual: Ritual is an emerging decentralized AI infrastructure project building a network for AI model hosting and inference with on-chain primitives. It overlaps with Macrocosmos's vision of decentralized AI infrastructure but focuses on inference rather than training or data.
- Filecoin: Filecoin is a decentralized storage network and DePIN pioneer. While storage-focused, it is comparable as an established Web3 decentralized infrastructure network with token incentives for participants, paralleling Macrocosmos's subnet incentive model.
- Modulus Labs: Modulus Labs builds zero-knowledge machine learning (zkML) infrastructure for verifiable AI. It is an adjacent decentralized AI player with thematic overlap on trustworthy and verifiable AI, complementary to Macrocosmos's "white box" AI positioning.
Others
- Opentensor Foundation (Bittensor): Opentensor Foundation stewards the Bittensor protocol on which Macrocosmos operates its three subnets. It is the foundational infrastructure provider and ecosystem steward; Macrocosmos is a tenant-operator on its network.
Broad incumbents
- CoreWeave: CoreWeave is a large-scale GPU cloud provider specializing in AI training and inference workloads. It is the broad incumbent Macrocosmos positions itself against on AI compute, though centralized rather than decentralized.
- Together AI: Together AI is a cloud platform for open and custom model training, fine-tuning, and inference. It is a broad incumbent alternative to Macrocosmos's decentralized training and inference ambitions for AI developers.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat4 records
Key risks6 records
Key highlights7 records
Customer concentration
Macrocosmos social profiles
Digital presenceMacrocosmos compliance and trust
Trust signalCompliance3 records
Macrocosmos financial estimates
Financial estimateRevenue estimate
Valuation estimate
Macrocosmos leadership team
Management profileNumber of profiles
Profiles6 records
Macrocosmos funding detail
Funding detailFunding overview
Funding rounds
Investors
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Macrocosmos M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about Macrocosmos
What does Macrocosmos do?
Macrocosmos builds and operates three decentralized AI subnets on the Bittensor network: Apex (SN1) runs incentive-driven competitions where miners solve algorithmic optimization problems; IOTA (SN9) provides a distributed pre-training architecture that runs across heterogeneous, unreliable hardware; and Data Universe / Gravity (SN13) collects and distributes real-time social media data at scale (55+ billion rows from X, Reddit, and YouTube). Customers access the platform via web apps, Python APIs, a desktop contribution app, and a curated data marketplace.
Is Macrocosmos a public or private company?
Macrocosmos is a private company. It is classified as founder individual operated bootstrapped and is currently operating.
When was Macrocosmos founded?
Macrocosmos was founded in 2024. It employs 11 to 50 people.
Where is Macrocosmos based?
Macrocosmos is headquartered in London, United Kingdom, in the Europe region.
How does Macrocosmos make money?
Five revenue lines are on record. Gravity Data Subscription Plans are the primary driver. The others are gravity Pay-as-you-go, gravity Marketplace, apex Competition Fees and bittensor Network Incentives.
Who are Macrocosmos's main competitors?
Direct peers on record are Render Network, Gensyn, Akash Network and io.net. Emerging players are Ritual, Filecoin and Modulus Labs. Opentensor Foundation (Bittensor) is listed as an others. Broad incumbents are CoreWeave and Together AI.
Does Macrocosmos have an API?
Yes. The Macrocosmos API provides programmatic access to Data Universe (Gravity) for launching scraping tasks and retrieving structured datasets. Developers use the GravityClient to query social media data from X (Twitter), Reddit, and YouTube by keyword, username, date range, and filters with engagement metrics. Data is delivered in CSV or parquet formats. API keys are required for authentication. Developer documentation is at docs.macrocosmos.ai/subnets/subnet-13-data-universe/readme.
What industry is Macrocosmos in?
Macrocosmos's product category is Decentralized AI Infrastructure. Its primary akta.pro industry code is FSAPALAC, Decentralized GPU/AI Compute Networks (training/inference, GPU renting, model hosting), with a secondary code of FSAPALAK, DePIN Orchestration, Scheduling & Middleware (resource discovery, job scheduling, workload orchestration). Its NAICS code is 518210 and its SIC code is 7370.