What Mozilla Data Collective does

Mozilla Data Collective (MDC) operates a data sharing platform and marketplace for artificial intelligence training data, founded in September 2025 as a mission-locked British company (company number 17054959) and incubated by Mozilla Foundation. The platform enables data contributors—including researchers, NGOs, language communities, broadcasters, and civic organizations—to publish, license, and distribute datasets while retaining ownership and control over terms, with three access models available: openly licensed (CC0, CC-BY), community-governed, or compensated. MDC serves two participant groups: data contributors seeking fair value exchange and provenance preservation, and data consumers (AI/ML researchers, model trainers, technical organizations) requiring ethically-sourced, diverse training data for underserved languages and modalities.

The platform is built on cloud storage infrastructure (S3/R2) with a REST API, Python SDK, and presigned URLs for programmatic dataset access, supporting resumable downloads via range requests and enforcing 30 downloads/day per organization rate limits. Supporting features include an AI-powered Data Assistant for natural-language dataset discovery (Alpha, launched May 2026), a Request to Access gating workflow with Public/Restricted/Private options, a Data Provider Analytics Portal, a Dataset Notification subscription feature, and an open-source Apache 2.0 speech-to-text fine-tuning blueprint for Whisper and MMS models. MDC also operates a competition platform powered by DrivenData, hosting the inaugural Lost In Transcription competition with $20,000 in prizes for speech recognition models on underrepresented languages (Indonesian-Javanese, Nahuatl-Spanish, Spanish-English). Dataset coverage spans 1,001+ assets across ASR, TTS, MT, NLP, CV, and multimodal modalities, with global geographic reach including African, South Asian, European, and Latin American language communities.

MDC's business model is a marketplace take rate: data uploaders set their own prices for compensated datasets (14 listings among 1,001+ total), and MDC retains 5% per transaction to cover infrastructure costs while 100% of license fees flow to contributors. Payment processing is handled by Stripe. The go-to-market is community-led, leveraging Mozilla Foundation's existing networks, developer relations, conference presence (PyConAU 2025, Mozilla Festival Barcelona), and AI model competitions to attract both contributors and consumers. Distribution is primarily self-serve via the web platform, REST API, and a dedicated competitions site. No external funding rounds are disclosed; the company operates under Mozilla Foundation incubation with stated ambition for self-sustainability through platform fees rather than grant dependence.

Name

Mozilla Data Collective

Legal name

Mozilla Data Collective LTD

Website

https://mozilladatacollective.com

Company type

Private

Founded year

2025

Operating status

Operating

Short description

Mozilla Data Collective operates a data sharing platform and marketplace for AI training data, enabling researchers, NGOs, and language communities to publish and license datasets while retaining ownership. A 5% transaction fee on compensated datasets funds the mission-locked, Mozilla Foundation-incubated entity.

Ownership category

akta.pro rank

Product category

AI Data Marketplace

NAICS

Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (51821)

SIC

Services-Computer Programming, Data Processing, Etc. (7370)

akta.pro primary industry

Dataset Discovery, Marketplaces & Licensing (HDAAALAE)

akta.pro secondary industries

Data Sharing, Data Exchange & Data Marketplace Platforms (HDAEABAH), Data Pipelines for GenAI (Curation, Filtering, Deduplication, Copyright) (HDAAACAJ), Data Sharing & Data Marketplace Governance (HDAEADAK)

Keywords

• AI training data
• Dataset marketplace
• Speech recognition data
• Multilingual data platform
• Open data licensing

Headquarters

HQ city

London

HQ country

United Kingdom

HQ region

Europe

Offices1 record

Markets served

GTM type

B2B

Offering type

Digital Commerce or Content

Cost components

Technology or R&D, Infrastructure, Operations, Personnel, Marketing or Sales

Revenue model

1. Platform Fee on Dataset Transactions: MDC charges a 5% platform service fee on compensated dataset purchases, bundled into the total price shown to data consumers at checkout. This covers infrastructure, hosting, and support costs. The remaining 95% of dataset license fees goes directly to the data uploader.
2. Dataset Licensing Fees: Data providers can set prices for their datasets. All dataset fees (100%) go directly to uploaders. MDC acts as a payment facilitation intermediary, processing payments through Stripe. This revenue stream supports uploaders including communities with lived experience of colonialism who need value recognition for their labor.

Pricing tiers

Go-to-market motion3 records

Distribution channels4 records

Marketing channels7 records

Core offering

Mozilla Data Collective operates a cloud-based data sharing platform and marketplace for AI training datasets, enabling data contributors to upload, license, and distribute datasets under open, community-governed, or compensated models. Data consumers—including AI/ML researchers, developers, and organizations—discover and download multilingual, multicultural, and multimodal datasets through a web interface, REST API, and Python SDK. The platform charges a 5% fee on compensated dataset transactions, with 100% of dataset license fees going to contributors.

Product overview

Mozilla Data Collective is a data sharing platform and marketplace built around human data agency and fair value exchange. The core platform (mdc-platform) enables data uploaders to publish, license, and distribute datasets with full ownership and control, while data consumers discover and download datasets for AI/ML training. The platform supports three access models: openly licensed (CC0, CC-BY), community-governed, and compensated (paid access with MDC taking a 5% platform fee). Supporting products include a REST API and Python SDK for programmatic access, the speech-to-text-finetune open-source blueprint for Whisper/MMS model fine-tuning, a Data Assistant for AI-powered dataset search, a Data Provider Analytics Portal, an MDC Competitions platform powered by DrivenData, a Request to Access gating feature, and a Dataset Notification subscription feature.

Differentiator

Problem solved

Functional benefit

Products and services

• Mozilla Data Collective Platform Core cloud-based data sharing platform and marketplace enabling data uploaders to publish, license, and distribute AI training datasets while retaining ownership and control. Data consumers discover and download multilingual, multicultural, and multimodal datasets through a web interface. Supports three access models: openly licensed (CC0, CC-BY), community-governed, and compensated with a 5% platform fee.
• MDC Public API REST API enabling developers to programmatically discover, access, and download datasets from the MDC platform in any programming language. Base URL: dev.mozilladatacollective.com/api. Supports GET /datasets/:datasetId and POST /datasets/:datasetId/download endpoints with presigned URL downloads.
• MDC Python SDK Python software development kit providing convenient access to MDC platform datasets. Enables dataset loading, authentication, and download management for Python-based machine learning workflows.
• speech-to-text-finetune Blueprint Open-source blueprint enabling developers to fine-tune Whisper-based or MMS-based speech-to-text models using MDC-hosted datasets. Includes local transcription app, model comparison demo, and GitHub Codespaces/Google Colab instant run options.
• MDC Competitions AI model competition platform powered by DrivenData, hosting prize-based competitions for building models on underserved linguistic datasets. Currently features the Lost In Transcription competition for speech recognition with $20,000 in prizes across Indonesian-Javanese, Nahuatl-Spanish, and Spanish-English contexts.

Quantifiable outcome

• 100% of dataset fees go directly to data contributors
• +2 more outcomes

Named customers6 records

Segments4 records

Ideal customer profiles4 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

Integration2 records

AI capability9 records

Feature6 records

Partnerships

Ten partnerships are on record, tiered core and minor.

• Mozilla FoundationcoreStrategic or Co-development Partner · 1 September 2025Mozilla Data Collective is incubated by Mozilla Foundation and backed by Mozilla.org. MDC is a mission-locked British company governed by Mozilla Foundation to ensure the platform's purpose is preserved. This relationship provides MDC with Mozilla's community networks, brand credibility, and organizational infrastructure.
• CLARIN (Common Language Resources and Technology Infrastructure)minorStrategic or Co-development PartnerMozilla Data Collective datasets are now discoverable through CLARIN's Virtual Language Observatory, expanding visibility for community-governed language datasets and improving exploration for European researchers, developers, and language technology practitioners.
• DrivenDatacoreTechnology or IntegrationMDC Competitions platform is powered by DrivenData (competitions.mozilladatacollective.com). This provides the technical infrastructure for running AI model competitions with prize components.
• Common VoicecoreStrategic or Co-development PartnerMozilla Data Collective hosts segments and derivatives of Mozilla Common Voice datasets (CC0-licensed speech data). Common Voice launched at Mozilla Foundation in 2017. MDC provides stewardship support including respecting right to be forgotten and maintaining versioning.
• Institute of African Digital HumanitiesminorStrategic or Co-development PartnerContributes African language datasets including TTS datasets, sociocultural datasets, and ALCAM multimodal datasets for underrepresented African languages (Bafia, Mada, Suundi, Foulbe, etc.).
• AmnesiaminorStrategic or Co-development PartnerMexican civil society organization preserving public information. Contributes 40+ datasets documenting freedom of information filings across Mexican states, addressing digital amnesia caused by government transitions.
• PROXIMAminorStrategic or Co-development PartnerContributes Urdu and Sindhi NLP datasets for LLM training including instruction tuning datasets, synthetic datasets, and specialized domain datasets (medical, legal, financial).
• Radio Free Europe/Radio Liberty (RFE/RL)minorStrategic or Co-development PartnerInternational broadcaster reaching 44 million people weekly across 18 countries in 24 languages, contributing datasets to serve communities through AI-ready language data.
• Open Home FoundationminorStrategic or Co-development PartnerContributes TTS datasets to enable voice assistants to speak underserved languages, supporting open source voice assistant development beyond mainstream languages.
• StripecoreTechnology or IntegrationPayment processor for dataset transactions. Stripe facilitates distribution of dataset fees to uploaders and platform fees to MDC, enabling the compensated datasets feature.

Scale indicators4 records

Recent moves8 records

Expansion highlights6 records

Direct peers

• Hugging Face: Hugging Face operates the largest community hub for AI datasets and models, including a Datasets library and Spaces for dataset hosting — directly overlapping MDC's dataset discovery and hosting core.
• Kaggle: Kaggle Datasets is a free dataset marketplace widely used by AI/ML practitioners; serves the same dataset discovery and consumption workflow as MDC, with a much larger user base.
• Roboflow: Roboflow hosts a large catalog of computer vision datasets with discovery and tooling; overlaps MDC's CV dataset hosting and developer-focused dataset distribution.

Broad incumbents

• AWS Data Exchange: AWS Data Exchange is a cloud-native marketplace for third-party data, including AI training datasets; competes for enterprise buyers seeking licensed datasets with managed billing and distribution.
• Scale AI: Scale AI is a major AI data platform spanning labeling and dataset curation across modalities; competes for the same AI training data customers MDC serves, with significant enterprise sales reach.
• Appen: Appen is an established AI training data provider with global crowd contributors and multilingual speech/language datasets; competes for the same model-training data sourcing budgets.

Emerging players

• Common Crawl: Common Crawl is a nonprofit providing openly licensed web crawl data widely used in LLM training; an adjacent peer in ethical/open data supply, though focused on web text rather than community datasets.
• Surge AI: Surge AI operates an AI data labeling and dataset platform focused on RLHF and frontier model evaluations; competes for premium AI data buyers with a contributor-friendly positioning.
• Snorkel AI: Snorkel AI provides data-centric AI development tooling including programmatic labeling and dataset management; serves a similar AI builder audience with adjacent data platform capabilities.

Others

• Open Data Institute: The Open Data Institute advocates for and curates open, ethical data ecosystems; thematically aligned with MDC's mission and a potential policy/distribution partner in the trusted data commons space.

Market position

Strengths4 records

Weaknesses5 records

Competitive moat5 records

Key risks6 records

Key highlights6 records

Customer concentration

• LinkedInlinkedin.com/company/mozilla-data-collective

Compliance2 records

Revenue estimate

Valuation estimate

Number of profiles

Funding overview

Funding rounds1 record

Investors1 record

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Mozilla Data Collective