Probabl and Intel today announced a collaboration to optimize and advance scikit-learn, the world's most widely used open-source library for machine learning, so that data scientists and enterprises get the best possible performance and total cost of ownership (TCO) from Intel hardware, throughout a heterogeneous compute landscape, from Xeon CPUs to Intel GPUs and beyond - across AI and machine learning workloads.
Unveiled at VivaTech in Paris, the collaboration brings together Probabl, the official steward of scikit-learn, and Intel, working hand in hand to make scikit-learn run faster and more efficiently on Intel platforms, with a particular focus on data center environments where performance and cost-efficiency at scale matter most.
The initiative is built on a shared commitment to open source. Improvements developed through this work are intended to be contributed back to the scikit-learn project under its permissive BSD license, benefiting the entire community rather than any single vendor. The goal is straightforward: better out-of-the-box performance for the millions of users who rely on scikit-learn every day, and a clear path for organizations to maximize the value of their Intel-based infrastructure for AI and machine learning.
"Enterprises run scikit-learn at scale while performance and cost directly shape what they can build. Our work with Intel is about delivering measurable gains in both, openly and for everyone, so the whole community moves forward together." said Yann Lechelle, Executive President, Probabl.
By aligning Probabl's core-expertise in scikit-learn with Intel's leadership in compute hardware, this joint effort aims to set a new standard for performance and efficiency in everyday machine learning, from individual practitioners to large-scale data center deployments.
Further details on the scope and roadmap of this program will be shared in the coming weeks.
Probabl is the tabular AI company founded by the creators of scikit-learn. An INRIA spin-off, Probabl was built by the core scikit-learn team – the engineers who have maintained and evolved the library for over 15 years. By June 2026, scikit-learn had been downloaded over 4.8 billion times - more than 2 billion times in the last 12 months alone. Probabl also builds Skore, the next-generation data science platform that enables teams to leverage the latest AI technologies, including agentic and tabular foundation models, with the scientific guardrails that professional client delivery demands and produce fully auditable and trustable data science deliverables. Learn more at Probabl.ai.
Intel is a global leader in the design and manufacturing of computing and AI infrastructure, powering data centers, cloud platforms, and devices worldwide. Intel is committed to advancing an open, performant, and accessible AI software ecosystem. Learn more at intel.com.
This release announces the general terms of the collaboration. Specific technical and commercial details are not disclosed at this time.