Domyn tests 1T-parameter MoE training on JUPITER supercomputer
11/15/2026 — 11/15, 23:17·1 sources (all)·1 reports (all)
Key takeaways
- Domyn published a technical post, also carried on the PyTorch blog, describing an experiment in which the EUROPA consortium trained a 1T-par…
Background & analysis
On June 19, 2026, the EUROPA consortium was selected by the European Commission as the winner of the EuroHPC Frontier AI Grand Challenge. The consortium is led by Domyn, a Milan-based company. On November 12, 2026, Domyn published a technical post describing training trials for a 1 trillion parameter mixture-of-experts (MoE) model on the JUPITER supercomputer. The same write-up appeared on the PyTorch blog that same day.
According to the post, the trials were carried out on the JUPITER supercomputer located in Jülich, Germany. They reached a throughput of 1300 tokens per second per GPU, and the post explicitly states that these are trials and not a completed pretraining run. The performance gains are attributed to three engineering changes, one of which is splitting work by whole experts to work alongside the Muon optimizer.
As of the available reporting, the public record stops at the disclosure of this experimental training effort. There is no information yet about a completed pretraining run or any further progression beyond the trial described in the post.
AI-generated from 1 reports · updated 2 days ago
- Heat index
- 34
- All sources
- 1
- All reports
- 1
- First seen
- 2 days ago
Reports on this story headlines open the original
A technical post from Domyn, also published on the PyTorch blog, describes training trials of a 1T-parameter mixture-of-experts model on the JUPITER supercomputer at Forschungszentrum Jülich, reaching 1,300 tokens per second per GPU. The company says these are trials, not a completed pretraining run, and credits three engineering changes, including sharding experts whole to work with the Muon optimizer.
DEV Community · AIAI score 59
Other stories people are talking about
- Heat index 1345RisingNvidia board not discussing Jensen Huang's return, probe into his conduct advances22 sources
- Heat index 1241SurgeCalifornia Sues OpenAI Over Agents Escaping Tests, Intruding into Hugging Face15 sources
- Heat index 901Alibaba releases MoE model Qwen4-Max with 1M token context for low-cost, long-running tasks17 sources
- Heat index 695US Commerce Department issues interim final rule restricting remote access to AI chips via overseas data centers11 sources
- Heat index 532Anthropic to begin trading on Nasdaq this week under ticker ANT9 sources
How is heat calculated?About the methodHide
Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.
This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.
- Surge
- Discussion rising fast
- New
- First report within 6 hours
- Rising
- Still gathering discussion

