Substra Foundation

Name

Substra Foundation

Legal name

Substra

Website

https://substra.ai

Company type

Private

Founded year

2019

Operating status

Operating

Headcount range

1–10 employees

Short description

Substra Foundation hosts an open-source federated learning platform that enables healthcare institutions and pharmaceutical companies to train ML models on distributed datasets without sharing raw data.

Ownership category

akta.pro rank

Headquarters

HQ city

Nantes

HQ country

France

HQ region

Europe

Offices1 record

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Personnel, Infrastructure, Operations

Revenue model

1. Open Source Framework: Substra is an open-source project hosted by the Linux Foundation for AI and Data. Revenue model is not explicitly disclosed in available sources. The project appears to be community-supported with Owkin as the primary contributor.

Go-to-market motion1 record

Distribution channels1 record

Marketing channels4 records

Core offering

Substra is an open-source federated learning software that enables the training and validation of machine learning models on distributed datasets without sharing raw data. It provides a flexible Python interface and a web application to run federated learning at scale, deployed in production by hospitals and biotech companies.

Product overview

Substra is an open source federated learning platform hosted by the Linux Foundation for AI and Data. The platform consists of a core product (Substra) with multiple modular components: substrafl provides a high-level Python library for running complex federated learning experiments, while the low-level substra library handles network interaction. Supporting modules include substra-backend, substra-frontend, orchestrator (task orchestration), substra-tools (Python wrappers), substra-documentation, charts (deployment), and substra-tests. Originally developed by Owkin, the platform enables hospitals and biotech companies to train and validate ML models on distributed datasets without sharing raw data.

Differentiator

Problem solved

Functional benefit

Products and services

• Substra Open-source federated learning software that enables training and validation of machine learning models on distributed datasets without sharing raw data. Provides a flexible Python interface and a web application to run federated learning at scale. Designed for hospitals, biotech, and pharmaceutical companies operating on sensitive, privacy-constrained data.
• SubstraFL A high-level federated learning Python library used to run complex federated learning experiments at scale on a Substra network. Targeted at data scientists and researchers building FL workflows.
• Substra Python Library Low-level Python library used to interact with a Substra network for federated learning operations. Provides programmatic control over network interactions for developers.

Quantifiable outcome

• First-ever federated deep learning model trained on distributed histology images stored behind hospital firewalls
• +2 more outcomes

Named customers3 records

Segments3 records

Ideal customer profiles3 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

AI capability3 records

Feature4 records

Partnerships

Eight partnerships are on record, tiered core, key and secondary.

• OwkincoreStrategic or Co-development PartnerOwkin originally developed Substra and remains the main contributor to the project. Owkin coordinates federated learning initiatives including MELLODDY and has been instrumental in Substra's evolution.
• Linux Foundation for AI and Data (LFAI)coreStrategic or Co-development PartnerLinux Foundation for AI and Data (LFAI) hosts Substra, providing governance, infrastructure, and institutional backing for the open-source project.
• NVIDIAkeyTechnology or IntegrationNVIDIA is a technology partner in the MELLODDY project, contributing GPU infrastructure and expertise to the federated learning platform for drug discovery.
• AWSkeyTechnology or IntegrationAWS provides cloud infrastructure for MELLODDY, deployed over an AWS private cloud to support federated learning across pharmaceutical companies.
• KU LeuvensecondaryTechnology or IntegrationAcademic research partner in the MELLODDY consortium, contributing research expertise to the federated learning platform.
• IktossecondaryTechnology or IntegrationAI drug discovery company partnering in MELLODDY project to contribute machine learning expertise for predictive model training.
• LoodsesecondaryTechnology or IntegrationTechnology partner in MELLODDY contributing infrastructure and orchestration expertise for the federated learning platform.
• BMEsecondaryTechnology or IntegrationPartner in the MELLODDY consortium providing expertise and resources to the federated learning platform for pharmaceutical research.

Scale indicators5 records

Recent moves6 records

Expansion highlights4 records

Emerging players

• FedML: Open-source and commercial federated learning platform supporting distributed training across cloud, edge, and on-device environments. Comparable as an FL tooling provider targeting both research and enterprise adoption.
• Sherpa.ai: Federated learning and privacy-preserving AI platform with applications across healthcare and regulated industries. Comparable as a commercial FL platform addressing similar data collaboration use cases.
• PySyft (OpenMined): Open-source library for privacy-preserving machine learning including federated learning and secure computation. Comparable as an open-source toolkit enabling collaborative ML on sensitive data without centralized sharing.

Broad incumbents

• Owkin: Commercial AI-for-health company that originally developed Substra and remains its main contributor. Comparable as a provider of federated learning solutions for hospitals and pharmaceutical companies, but operating as a commercial SaaS rather than an open-source foundation project.
• NVIDIA FLARE: NVIDIA's federated learning SDK for distributed AI training, particularly in healthcare. Comparable as a production-grade FL platform, but backed by NVIDIA's broader AI infrastructure portfolio rather than foundation governance.
• TensorFlow Federated (Google): Google's open-source framework for federated learning and computations. Comparable as an open-source FL toolkit, though embedded within Google's broader ML ecosystem rather than governed as a multi-vendor foundation project.

Direct peers

• OpenFL (Intel): Intel-developed open-source framework for federated learning in healthcare and life sciences. Directly comparable as a vendor-backed FL toolkit targeting clinical and research use cases similar to Substra.
• Flower (Adap): Open-source federated learning framework widely adopted in research and industry. Directly comparable as a vendor-neutral FL toolkit serving ML researchers and enterprises building privacy-preserving distributed training systems.
• FATE (Federated AI Technology Enabler): Linux Foundation-hosted open-source federated learning project. Closely comparable as a foundation-governed FL framework, though its emphasis on industrial/financial applications differs from Substra's healthcare focus.

Others

• Linux Foundation AI & Data Foundation: Umbrella foundation that hosts Substra and other open-source AI/data projects. Comparable as the governing entity providing institutional infrastructure, governance, and ecosystem support for foundation-hosted AI initiatives.

Market position

Strengths4 records

Weaknesses4 records

Competitive moat5 records

Key risks5 records

Key highlights6 records

Customer concentration

• GitHubgithub.com/substra
• Xx.com/Substra_org

Revenue estimate

Valuation estimate

Number of profiles

Funding overview

Funding rounds

Investors

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 Substra Foundation