Federated learning
Federated learning can improve models without pooling raw user data, and on-device inference can work alongside it. Google's documented deployments combine federated methods with privacy controls. The useful test is how much a model improves within a set privacy and communication budget. Whether a sector is legally barred from centralising data matters less.
Research snapshot · updated Nov 12, 2026
Core thesis
FL trains models across institutions without sharing raw data. Healthcare (hospital networks), finance (cross-bank fraud detection), and mobile (on-device learning) are primary use cases. But FL adds significant complexity, reduces model performance vs centralised training, and faces communication overhead challenges. Regulatory drivers (GDPR, HIPAA) push adoption; engineering complexity holds it back.
State of the art (2026)
Federated learning improves a model using data held on many devices or in many organisations, without collecting the raw data in one place. Google's documented deployments combine it with other privacy controls. Two open-source frameworks are widely used, NVIDIA FLARE and Flower. Flower also sells an enterprise product meant for production use, with security, scale and support. That is the supplier's own description, and we have no independent evidence on how those deployments perform. It remains open whether federated learning stays a category of its own or becomes a feature of on-device models, synthetic data and data clean rooms.
Explore the deeper analysis in CanaryIQ
Explore what is covered below. These previews show the structure; log in to read the full analysis.
Signal stack
Evidence stacked leading → lagging
10 signals
Signal stack
Evidence stacked leading → lagging
Technology-native KPIs
Metrics that predict trajectory, tracked over time
4 tracked
Technology-native KPIs
Metrics that predict trajectory, tracked over time
Landscape map
Who builds what, and who depends on whom
95 players · 6 layers
Landscape map
Who builds what, and who depends on whom
Catalyst calendar
Dated events that will move the position
5 ahead
Catalyst calendar
Dated events that will move the position
Technology roadmap
Milestones on the path to maturity
8 milestones
Technology roadmap
Milestones on the path to maturity
Watchlists
Companies, people and papers, each with a condition for removing it
20 · 20
Watchlists
Companies, people and papers, each with a condition for removing it
Decision frameworks
What our view means for your job
Locked
Decision frameworks
What our view means for your job
Thesis changelog
When our view changed, and why
8 updates
Thesis changelog
When our view changed, and why
Change our mind
2 disconfirming conditions
You've read the verdict. The file is much deeper.
The full signal stack, technology-native KPIs tracked over time, the landscape of who depends on whom, the dated catalyst calendar, decision frameworks for every desk, live watchlists and a record of every time our view on Federated learning has changed. All of it is in CanaryIQ.