Details
- Status
- Open
- Reference
- AI FOR INDUSTRIAL INNOVATION - LARGE SCALE ACCESS
- Publication date
- 9 April 2025
- Opening date
- Deadline model
- Multiple cut-off
- Deadline dates
- 30 Apr 2025, 10:00 / 15 May 2025, 10:00 / 30 May 2025, 10:00 / 16 Jun 2025, 10:00 / 30 Jun 2025, 10:00 / 15 Jul 2025, 10:00 / 31 Jul 2025, 10:00 / 14 Aug 2025, 10:00 / 1 Sep 2025, 10:00 / 15 Sep 2025, 10:00 / 30 Sep 2025, 10:00 / 15 Oct 2025, 10:00 / 31 Oct 2025, 10:00 / 14 Nov 2025, 10:00 / 1 Dec 2025, 10:00 / 15 Dec 2025, 10:00 / 5 Jan 2026, 10:00 (CET)
Description
This access call is intended:
- For Industry users
- For More than 50,000 GPU hours
- To provide Approval within 10 working days
Large Scale Access call is intended for applications performing AI activities with high-impact, high-gain innovative research, justifying the need for and the capacity to use large allocations in terms of compute time, data storage and support services. The allocations are granted for a period of three, six, or twelve months, depending on the duration needs.
Access is granted within 10 working days from cut-off date provided that the proposal successfully passes the technical and peer-review evaluation organised by the EuroHPC JU Peer-review office.
The Large Access call is continuously open with predefined cut-off dates (two per month).
AI Factories Systems
The supercomputer systems and their services provided for the Fast Lane Access call by EuroHPC Hosting Entities are:
(*The above EuroHPC systems are sorted in order to show the last available system that entered production. **Maximum resources available per month. ***Maximum and minimum resources available per proposal, overall.) All resources in the table are displayed in GPU hours.
Explanation: Proposals with very large needs in terms of GPU time should consider the maximum amount available per month for the targeted partition when submitting their proposal. Example: If an application request the maximum amount available per proposal of a EuroHPC system (for example, MN5 ACC = 4 066 000 GPU hours), its duration must be divided by the maximum GPU time available per month (for example, MN5 ACC = 336 000 GPU hours). In this specific case, the minimum duration of the project would be 12 months (Request / maximum GPU available per month = minimum duration of the project).
Explore AI Factories FAQs Back to overview of Access Modes