On October 15, 2026, European physics research center CERN and Mistral AI signed a statement of intent to build a Large Physics Model trained on billions of simulated particle collisions. The model will be the first foundation model for collider physics, replacing specialized particle-reconstruction algorithms. CERN currently holds one exabyte of Large Hadron Collider data and anticipates a tenfold increase in collision volume after a 2030 upgrade. No timeline or budget was disclosed.
Announcement of the CERN and Mistral AI partnership
- ▪On October 15, 2026, European physics research center CERN and Mistral AI signed a statement of intent to establish a long-term strategic partnership to develop foundation models for scientific discovery
- ▪European physics research center CERN and Mistral AI disclosed no timeline or budget details when announcing their strategic artificial intelligence partnership on October 15, 2026
Goals and scope of the Large Physics Model
- ▪CERN and Mistral AI aim to build a Large Physics Model trained on billions of simulated particle collisions to replace specialized algorithms currently used to reconstruct and identify particles
- ▪CERN stated that the planned Large Physics Model will be the first foundation model developed specifically for collider physics
- ▪CERN and Mistral AI stated that foundation models created through their partnership could eventually be adapted for applications in other scientific fields beyond physics
Data volume and future capacity at the Large Hadron Collider
- ▪CERN expects the volume of particle collisions at the Large Hadron Collider to increase tenfold following an ongoing facility upgrade planned for around 2030
- ▪CERN holds approximately one exabyte, equivalent to one million terabytes, of experimental data generated by the Large Hadron Collider
Story comments
Loading comments…