Flower AI Day is set to be a landmark for the open-source AI movement -- featuring keynotes from world-leading engineers and scientists, ground-breaking demos of brand-new Flower technology, and interactive expert-led panels. We are gathering together due to a fast evolving change in the AI landscape that will impact everything: Open-source Decentralized Super Intelligence has arrived.

Flower AI Day will explore the rapid arrival of  open-source decentralized superintelligence. Core relevant sub-topics include:

on-device LLMsprivacy-preserving technologies (PETs)decentralized AIfederated learningcompliance-ready AIconfidential computeenterprise-scale fed AI deployments

Featured speakers

More speakers coming soon

Tony James

Chief Architect, Science and Space
Red Hat

Sara Hooker

VP Research, Head of Cohere Labs
Cohere

Sudhir Upadhyay

Head of Engineering
Kinexys by J.P. Morgan

Ce Zhang

CTO at Together AI
Together AI

Joseph Spisak

Product Director, Generative AI
Meta

Timothy Chou

Lecturer/ Founder Pediatric Moonshot
Stanford

Daniel J. Beutel

Co-Creator of Flower
Flower

Nicholas Lane

Co-Founder & CSO
Flower
Cambridge
Stay Tuned for Surprise Speakers

Preliminary Agenda

1:00 PM

☕️ Registration and Reception

Check in for badge and swag. Network over coffee.

1:30 PM

🌼 Welcome and Flower Keynote

Early preview of new Flower tech and demos including AI using federations, decentralization, privacy tech, and on-device compute.

2:00 PM

🎙️ The Road to Open-source Decentralized AGI

Talks on decentralized training that scale to frontier-quality foundation models.

2:45 PM

💬 Fireside Chat: Open Source AI -- Going All In

3:15 PM

☕️ Refreshment Break and Networking

3:45 PM

🎙️ Enterprise-scale Federated AI: A View from the Front-line

Real deployments, lessons learned, successes, and emerging patterns inside large organizations.

4:30 PM

🎙️ The Power of On-device AI and Confidential Compute for Privacy and Compliance

Speakers cover solutions to privacy, regulation and safety using local AI and/or confidential compute.

5:00 PM

☕️ Refreshment Break and Networking

5:30 PM

🎙️ Enabling Universal AI: Every Language, Every Location, Every Application Domain

Unlocking overlooked AI applications and neglected users/regions by improving training data coverage and model eval.

6:15 PM

💬 Panel: Federated AI is Here. What Happens Next?

6:45 PM

🍷 Buffet Dinner & Open Bar Reception

Get your tickets now

Join us for our next event

Decentralized AI Hackathon 2025

SEPT. 26
Stanford University

Flower Labs is bringing builders, researchers, and the open-source community together at Stanford University for a full day of hands-on building, lightning talks, and demos—shaping the future of decentralized intelligence. Participation is free, but spots are limited, register today.

What to expect at Flower AI Day? Check out Flower AI Summit from early 2025.

Catch the best of Flower AI Summit 2025, March 26–27 in London — don’t miss the announcements of Flower Intelligence and COLLECTIVE-1 and talks from Mozilla, Databricks, Vana, Owkin, Hugging Face, Intel and Nvidia. Now available online.

Fireside Chat with Thomas Wolf
Flower LabsSubscribe

Fireside Chat with Thomas Wolf

Thomas Wolf · Co-Founder and CSO of HuggingFace

User-Owned Foundation Models
Flower LabsSubscribe

User-Owned Foundation Models

Anna Kazlauskas · creator of Vana and CEO of Open Data Labs

FL with IP Protection of Flower Models using NVIDIA FLARE
Flower LabsSubscribe

FL with IP Protection of Flower Models using NVIDIA FLARE

Holger Roth · Principal FL Scientist at Nvidia

FL for Medical Image Analysis
Flower LabsSubscribe

FL for Medical Image Analysis

Pramit Saha · DPhil Student at University of Oxford

Flower Federated AI: Ready to Deploy
Flower LabsSubscribe

Flower Federated AI: Ready to Deploy

Chong Shen Ng · Research Engineer at Flower Labs

Federated Foundation Model for Generalisable Disease Detection
Flower LabsSubscribe

Federated Foundation Model for Generalisable Disease Detection

Eden Ruffel · PhD Student at UCL

TAO: Fine-tuning without labels
Flower LabsSubscribe

TAO: Fine-tuning without labels

Jonathan Frankle · Chief Scientist at Databricks