OpenMined

What OpenMined does

OpenMined Foundation is a 501(c)(3) non-profit founded in 2017 and headquartered in the United States with distributed operations including the United Kingdom. It develops an open-source portfolio of privacy-enhancing technologies (PETs) — federated learning, secure multi-party computation, differential privacy, homomorphic encryption, and zero-knowledge proofs — packaged into PySyft (remote data science via Datasites), SyftBox (decentralized computation protocol, beta), SyftHub and Syft Space (network directory and decentralized RAG nodes forming the 'Protocol for Collective Intelligence'), syft-flwr (federated learning infrastructure on the Flower framework), BioVault (biomedical data-visitation platform), and PipelineDP (differential privacy pipelines co-developed with Google for Apache Spark and Apache Beam). The organization operates a federated AI network in which data remains under owner control while computation, queries, and attribution are coordinated across silos, with dedicated subnets serving Publishers, Genomics and Biomedical Researchers, Content Creators, and AI Auditors and Safety Researchers. OpenMined funds itself through donations and grants rather than commercial product sales: Stripe-collected one-time and monthly donations ($10-$1,000 suggested), a $500,000 Microsoft commitment for the Christchurch Call Initiative (2022), a $250,000 PyTorch/RAAIS fellowship grant (2019), and shared winnings from the $1.6M UK-US PETs Prize Challenges (2022). GTM is community-led via a 17,000+ member Slack community, GitHub distribution, and 370+ technical blog articles, layered with enterprise field-sales-style partnerships with Twitter/X, Reddit, LinkedIn, Dailymotion, Microsoft, Meta, Google, Anthropic, and government agencies including US Census Bureau, Statistics Canada, Istat, UN Statistics Division, NSF, and UK AI Safety Institute. Customers served include publishers and creators needing attribution for AI consumption of their content, biomedical consortia, social media platforms seeking privacy-preserving algorithmic transparency, and statistical agencies conducting cross-border data joins (e.g., 813 matching records in a US-StatCan join).

Name

OpenMined

Legal name

OpenMined Foundation

Website

https://openmined.org

Company type

Private

Founded year

2017

Operating status

Operating

Headcount range

1–10 employees

Short description

OpenMined is a 501(c)(3) non-profit that builds open-source privacy-preserving AI infrastructure — federated learning, differential privacy, secure computation — enabling organizations to query and analyze siloed data without centralizing it, serving publishers, biomedical researchers, AI auditors, and statistical agencies.

Ownership category

akta.pro rank

Product category

Privacy-Enhancing Technologies

NAICS

Software Publishers (5132), Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518210)

SIC

Services-Prepackaged Software (7372), Services-Computer Programming, Data Processing, Etc. (7370), Services-Computer Integrated Systems Design (7373)

akta.pro primary industry

Confidential AI & Privacy-Preserving ML (federated learning, MPC, HE, TEEs) (HDAAAKAI)

akta.pro secondary industries

Privacy-Preserving On-Device ML (secure enclaves, TEEs, differential privacy) (HDAAAJAH), Privacy-Preserving ML & Data Protection (e.g., federated learning, differential privacy) (HDAAAMAD)

Keywords

• Privacy-preserving computation
• Federated learning infrastructure
• Secure multi-party computation
• Differential privacy frameworks
• Open-source AI tools

Headquarters

HQ city

Oxford

HQ country

United Kingdom

HQ region

Europe

Offices2 records

Markets served

GTM type

B2B

Offering type

Software

Cost components

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

Revenue model

1. Donations and Grants: As a 501(c)(3) non-profit foundation, OpenMined operates primarily through donations, grants, and partnership funding. The organization accepts one-time and monthly donations via Stripe, with suggested amounts ranging from $10 to $1,000. No commercial revenue streams from product licensing or subscriptions are indicated.

Pricing tiers

Go-to-market motion2 records

Distribution channels5 records

Marketing channels7 records

Core offering

OpenMined develops and distributes open-source privacy-preserving technologies (PETs) that enable secure computation across siloed data without centralizing sensitive information. Its product portfolio includes PySyft for remote data science, SyftBox for decentralized privacy-preserving computation, SyftHub and Syft Space for federated AI-powered knowledge sharing, syft-flwr for federated learning infrastructure, BioVault for biomedical research, and PipelineDP for differential privacy pipelines. The products serve enterprises, governments, researchers, publishers, and content creators needing to derive insights from data while maintaining privacy, attribution, and control.

Product overview

OpenMined offers a portfolio of privacy-preserving AI products organized around its federated AI network vision. The core offerings include SyftHub (network directory/registry) and Syft Space (decentralized knowledge node for creating queryable RAG endpoints), which together form the Protocol for Collective Intelligence. Supporting products include syft-flwr (federated learning infrastructure built on Flower), SyftBox (decentralized computation protocol in beta), and PySyft (remote data science with Datasite model). BioVault provides domain-specific biomedical research capabilities. The ecosystem is complemented by Attribution-Based Control as a conceptual framework, PipelineDP (differential privacy pipelines with Google partnership), and developer tools including Python/TypeScript SDKs and Syft MCP for AI model integration. OpenMined is a 501(c)(3) non-profit foundation building the public network for non-public information.

Differentiator

Problem solved

Functional benefit

Brands

• SyftHub: Directory for Collective Intelligence - a registry mapping who has knowledge, what they're willing to share, and how to reach them for decentralized knowledge sharing.
• Syft Space
• SyftBox
• syft-flwr
• PySyft
• BioVault

Products and services

• SyftHub Network directory and registry for OpenMined's collective intelligence network that maps who has knowledge, what they are willing to share, and how to reach them. Enables discovery of knowledge endpoints across the network with federated RAG and Mixture of Experts capabilities for data scientists, researchers, builders, and AI application developers.
• Syft Space Decentralized node product that allows anyone with knowledge to share by uploading documents or connecting vector databases (Weaviate, Qdrant, ChromaDB), connecting AI models (OpenAI, Anthropic, Ollama, vLLM), and creating queryable RAG endpoints with full control over access, attribution, and pricing for content creators, publishers, and knowledge owners.
• syft-flwr Production-ready federated learning infrastructure built on the Flower framework providing governance, security, and collaboration tools for FL projects. Features include auto-discovery of data sites, built-in identity and access control, code approval workflows, and Docker/Kubernetes deployment for multi-hospital research consortia, academic collaborations, and financial services fraud detection teams.
• SyftBox Open-source beta protocol enabling developers and organizations to build, deploy, and federate privacy-preserving computations across a decentralized network without centralizing data. Data stays on owner infrastructure for developers and organizations building distributed privacy-preserving applications.
• PySyft Open-source technology for remote data science enabling data scientists to perform analyses on sensitive data without compromising confidentiality. Implements a Datasite model with code submission, manual code review workflows, and automated approvals using privacy-enhancing technologies for data scientists, ML researchers, and organizations handling sensitive data.
• BioVault Open-source platform for biomedical research enabling collaborative analysis across institutions without transferring sensitive data. Implements a data-visitation model where approved analyses travel to data rather than data moving to analysts, with mock datasets mirroring real structure for genomic, GWAS, single-cell, multi-omics, medical imaging, and clinical time-series studies for biomedical researchers and institutions.
• PipelineDP Open-source framework for applying differential privacy to large datasets using batch processing systems such as Apache Spark and Apache Beam. Co-developed with Google to provide production-level differential privacy implementation for data engineers and organizations handling large-scale sensitive datasets.

Quantifiable outcome

• Successfully completed international data joins between US Census Bureau, Statistics Canada, and Istat without exposing underlying data (813 matching records in US-StatCan join)
• +2 more outcomes

Named customers13 records

Segments10 records

Ideal customer profiles6 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

Integration10 records

AI capability12 records

Feature8 records

Partnerships

15 partnerships are on record, tiered flagship and core.

• AnthropicflagshipStrategic or Co-development Partner · 1 November 2024Pilot experiment using NVIDIA H100 secure enclaves and PySyft for AI evaluation across organizations. Anthropic provided an open-source model as proxy for testing while UK AISI provided confidential test datasets. Demonstrates practical feasibility of mutual privacy protection in AI auditing.
• White House OSTP/NSF/PCASTflagshipStrategic or Co-development Partner · 1 May 2024Invited participation in White House event recognizing NAIRR Pilot launch and PCAST AI report. One of 26 industry contributors meeting with government agencies, Congress, academia, and industry leaders.
• National Science Foundation (NSF)flagshipStrategic or Co-development Partner · 1 January 2024NAIRR Pilot launch partner providing US-based researchers access to compute credits, SyftBox software, and training for distributed privacy-preserving data science across consortia. Part of 25-organization partnership with NSF and 10 federal agencies.
• RedditcoreStrategic or Co-development Partner · 1 January 2023Partnership to help build Reddit for Research program enabling academic researchers to study online communities while protecting user privacy. OpenMined developing privacy-preserving infrastructure for secure data access and computation.
• DeepMindcoreStrategic or Co-development Partner · 1 November 2022Collaboration winning UK-US PETs Prize Challenge. Combined DeepMind's expertise with OpenMined's privacy-preserving technologies for demonstrating PET feasibility in real-world scenarios.
• UK AI Safety Institute (formerly Frontier AI Taskforce)flagshipStrategic or Co-development Partner · 1 October 2022Strategic partnership to develop and deploy technical infrastructure facilitating AI safety research across governments and AI research organizations. OpenMined is one of 11 organizations in the Taskforce's network of expertise in AI risks across biosecurity, cybersecurity, and deceptive behavior.
• Twitter (X)coreStrategic or Co-development Partner · 1 January 2022Partnership to test how PETs drive greater accountability by enabling ML research without sharing or exposing underlying data or models. Driven by Twitter's ML Ethics, Transparency and Accountability (META) team to enable third-party access to non-public Twitter data.
• GooglecoreStrategic or Co-development Partner · 1 January 2021Co-creation of PipelineDP, production-level tools for differential privacy. Partnership with Google's Anonymization team to build framework for applying differential privacy to large datasets using Apache Spark and Apache Beam.
• PyTorch (Meta)coreStrategic or Co-development Partner · 1 December 2019$250,000 investment through RAAIS Foundation for fellowship funding. Three fellowship categories: Core PySyft CrypTen Integration, Federated Learning on Mobile/Web/IoT, and Development Challenges. Partnership to combine PySyft and CrypTen for comprehensive PPML ecosystem.
• Facebook (Meta)coreStrategic or Co-development Partner · 1 May 2019Partnership to advance privacy-preserving AI through educational initiative. Facebook funded 5,000 scholarships for Secure and Private AI course on Udacity. OpenMined's PySyft was core technology in the course curriculum.
• US Census BureaucoreStrategic or Co-development PartnerParticipation in UN PET Lab initiative for international privacy-preserving data collaboration. Successfully demonstrated cross-border data joins with Statistics Canada and Istat without exposing underlying census data.
• Statistics CanadacoreStrategic or Co-development PartnerInternational collaboration through UN PET Lab demonstrating privacy-preserving statistical analysis. Successfully performed data joins with Istat and US Census Bureau using PySyft.
• Italian National Institute of Statistics (Istat)coreStrategic or Co-development PartnerPartnership through UN PET Lab for cross-border statistical collaboration. Demonstrated feasibility of international data joins while maintaining privacy protections.
• United Nations Statistics DivisioncoreStrategic or Co-development PartnerHosting UN PET Lab network gateway that connects national statistical organizations for privacy-preserving international collaboration using PySyft.
• Christchurch Call Initiative on Algorithmic OutcomesflagshipStrategic or Co-development PartnerMulti-stakeholder initiative including New Zealand, US governments, Twitter, Microsoft, and OpenMined to develop privacy-preserving research capabilities across multiple platforms for studying algorithmic outcomes.

Scale indicators9 records

Recent moves6 records

Expansion highlights6 records

Direct peers

• Flower Labs: Open-source federated learning framework (Flower) on which OpenMined's syft-flwr is built. Directly comparable in federated learning infrastructure and developer community positioning, but Flower Labs operates as a venture-backed commercial company while OpenMined is a non-profit.
• Duality Technologies: Provides secure collaboration on sensitive data using homomorphic encryption and PETs. Comparable in targeting regulated industries (healthcare, government, financial services) with privacy-preserving analytics.
• Cape Privacy: Privacy-preserving machine learning platform enabling encrypted collaborative analytics. Comparable in applying federated learning and cryptographic techniques to enterprise data collaboration use cases.
• TripleBlind: Commercial privacy-preserving data collaboration platform built around federated learning and cryptographic techniques. Directly comparable in cross-organization data collaboration value proposition, but operates a licensed enterprise model rather than open source.
• Enveil: Commercial enterprise product offering homomorphic encryption-based secure search and analytics. Directly comparable in delivering PET-powered enterprise solutions for cross-organization collaboration.
• Zama: Provides open-source homomorphic encryption and privacy-preserving ML tooling (Concrete-ML, TFHE-rs). Comparable as an open-source PET infrastructure provider targeting the same developer audience with overlapping cryptographic techniques.
• Inpher: Enterprise privacy-preserving analytics using secure multi-party computation and homomorphic encryption. Comparable in targeting financial services and government with cryptographic PET platforms.

Broad incumbents

• Google TensorFlow Federated: Google's open-source federated learning framework. Comparable as PET infrastructure, but distributed by a hyperscaler with vastly greater resources and proprietary extensions (e.g., PipelineDP partnership with OpenMined).
• Apple Private Cloud Compute: Apple's privacy-preserving AI inference architecture using secure enclaves. Comparable as a deployment of TEEs/PETs for confidential AI compute, but tightly integrated into Apple's ecosystem rather than a general-purpose platform.
• Microsoft SEAL / Azure Confidential Computing: Microsoft's homomorphic encryption library and secure enclave offerings. Comparable in providing PET building blocks at scale and is a direct OpenMined partner/funding source (Christchurch Call Initiative, $500K).

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks6 records

Key highlights7 records

Customer concentration

• Xx.com/openminedorg
• GitHubgithub.com/OpenMined
• LinkedInlinkedin.com/company/openmined

Compliance2 records

Revenue estimate

Valuation estimate

Number of profiles

Profiles1 record

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 OpenMined