The aim of the procurement is to conclude five framework agreements in lots 1 to 4 and in lot 6 (Section 21 VgV) for the modular procurement of high-performance (direct water-cooled) HPC and AI components, as well as a public contract for the procurement of a turnkey, AI-optimized HPC island solution (lot 5), which is suitable for data center operation and also approved by all hardware suppliers. The following must be offered for the individual lots:
Lot 1/LRZ_LXC: Direct water-cooled, accelerated HPC cluster
Lot 2/LRZ_GP: Direct water-cooled, unaccelerated HPC cluster
Lot 3/LRZ_BKI: Direct water-cooled, accelerated AI cluster
Lot 4/FAU_BKI: Direct water-cooled, accelerated AI cluster
Lot 5/FAU_CNT: Container-based, direct water-cooled, accelerated AI cluster
Lot 6/FAU_THR: Accelerated HPC cluster with mid-range GPUs

LOT-0001
Lot 1/LRZ_LXC.
Direct water-cooled system complex consisting of necessary racks, management servers, network components, and accelerated computing nodes with 2 CPUs with x86_64 compatible instruction set and 4 high-end GPUs per node. The computing nodes should be connected via InfiniBand.

LOT-0002
Lot 2/LRZ_GP.
Direct water-cooled system complex consisting of necessary racks, management servers, network components, and general-purpose computing nodes based on an x86_64 compatible instruction set. The computing nodes should be connected via InfiniBand.

LOT-0003
Lot 3/LRZ_BKI.
Direct water-cooled system complex consisting of necessary racks, management servers, network components, and AI computing nodes with at least 4 high-end GPUs per node. The computing nodes should be connected via InfiniBand.

LOT-0004
Lot 4/FAU_BKI.
Direct water-cooled system complex consisting of necessary racks, management servers, network components, and infrastructure to enable operation (hot water cooling infrastructure, power distribution, etc.), as well as AI computing nodes with at least 4 high-end GPUs per node. The computing nodes should be connected via InfiniBand.
In deviation from lots 1 to 3, the relevant NUTS code for the place of performance is DE252.

LOT-0005
Lot 5/FAU_CNT.
Turnkey, AI-optimized HPC island solution with approximately 1,200 GPUs and 10 PB NVMe storage, housed in "containers" for the installation of the entire system on open space and infrastructure to enable operation (hot water cooling infrastructure including cooling plant, power distribution, etc.). The accelerated computing nodes should contain at least 4 high-end GPUs and be connected via an InfiniBand network. The external connection is via 100 GbE trunks.
This is not a framework agreement, but a system supply contract.
In deviation from lots 1 to 3, the relevant NUTS code for the place of performance is DE252.

LOT-0006
Lot 6/FAU_THR.
Not necessarily directly water-cooled throughput-optimized GPGPU nodes with at least 4 mid-range GPUs (RTX PRO 6000 Blackwell or comparable). The computing nodes should only be connected via 25 Gbit Ethernet.
In deviation from lots 1 to 3, the relevant NUTS code for the place of performance is DE252.