Skip to content
Aperlena

TED - Tenders Electronic Daily (Publications Office of the EU) verified 3 weeks ago Contract award

Germany - Information systems and servers - AI cluster upgrade DGX v2

Original Deutschland - Informationssysteme und Server - KI-Cluster-Upgrade-DGX v2

Awarded Germany Oberbayern Software & IT systems

Buyer Universität der Bundeswehr München (UniBw M)

Key facts

Awarded
Procedure
Negotiated procedure The buyer selects candidates, then negotiates the offers before choosing. Full entry
Above the EU threshold A contract large enough that EU-wide rules apply and it must be published on TED. Full entry
View the official notice Watch this tender

Bidding is closed. An award or a correction on this notice: an email when the register sees it. Log in

Summary

Machine-generated summary

The Universität der Bundeswehr München (UniBw M) awarded a contract on 2025-11-27 for the delivery of an NVIDIA DGX B300 system for the dtec.bw project MissionLab and further research projects of the university. The buyer already operated a deep learning GPU cluster created through the contracts NU187, MT990 and LT218; contract NU187 was awarded on 16.12.2023 to DELTA Computer Products GmbH after a negotiated procedure without prior publication. That cluster comprised five DGX systems with 1,120 logical CPU cores, ten terabytes of RAM, 40 NVIDIA H100 Tensor Core GPUs, Infiniband networking and an NVLink-connected high-performance data server, used for training large language models, reinforcement learning and sensor evaluation. The new contract has been awarded.

Timing

  1. Awarded
  2. Published

Description

Auftragsgegenstand ist die Lieferung eines NVIDIA DGX B300 Systems für das dtec.bw Projekt MissionLab sowie weitere Forschungsprojekte der UniBw M. Auftragsgegenstand ist die Lieferung eines NVIDIA DGX B300 Systems. Für das dtec.bw Projekt MissionLab führte die UniBw M bereits im Jahr 2023 ein Vergabeverfahren für den Ausbau der vorhandenen Deep-Learning Infrastruktur durch. Um die installierte und etablierte Software, einschließlich des Trainings-Software-Frameworks, der Trainingsdaten und bereits vortrainierter neuronaler Netze, weiterverwenden zu können und eine gemein-same und effiziente Nutzung der bisherigen Deep-Learning Workstation zu gewährleisten, waren weitere DGX-H100 Systeme und die entsprechende NVIDIA-Netzwerk-Infrastruktur notwendig. Die UniBw M erteilte am 16.12.2023 nach Durchführung eines Verhandlungsverfahrens ohne Teilnahmewettbewerb ("Lieferung, Installation und Vernetzung eines Deep-Learning-Clusters (NU 187)") den Auftrag durch Zuschlagserteilung an die DELTA Computer Products GmbH (Delta). Aufgrund der Aufträge NU187, MT990, LT218 erhielt die UniBw M im Rahmen des Projekts MissionLab (dtec.bw) einen Deep-Learning-High-Performance-GPU-Cluster. Delta installierte und vernetzte fünf DGX-Systeme auf dem Gelände der UniBw M und integrierte einen High-Performance-/High-Storage-Server. Die UniBw M betreibt nun ein leistungsfähiges System mit 1.120 logischen Rechenkernen (CPUs), zehn Terrabyte Arbeitsspeicher (RAM), 40 NVIDIA H100 Tensor Core GPUs, einer Vernetzung der fünf Server über Infiniband und einem über NVLink angebundenen High-Performance-/High-Storage-Datenserver. Auch die 40 NVIDIA H100 Tensor Core GPUs (DGX-Systeme) sind über NVLink verbunden. Die UniBw M setzt ihren GPU-Cluster unter anderem für das Training von Large-Language-Modellen, die missions-bezogene Planung mittels Reinforcement Learning sowie für das Training von Sensoreinsatzmodellen und die automatisierte Sensorauswertung (Neuronale Netze, Vision Transformer) ein. Die Forschungsarbeiten im Projekt "MissionLab" erzeugen einen stetig wachsenden Bedarf an leistungsfähiger Recheninfrastruktur. Automatisierte Missionssysteme müssen in realitätsnahen Szenarien simuliert, bewertet und weiterentwickelt werden. Diese Untersuchungen basieren auf großen Datenmengen, komplexen KI-Modellen und iterativen Trainingsabläufen, die herkömmliche Serverstrukturen nicht mehr zuverlässig abbilden. Lange Rechenzeiten, begrenzter Grafikspeicher und fehlende Skalierbarkeit würden die Durchführung anspruchsvoller Simulationen sowie das Training moderner KI-Modelle erheblich verlangsamen und damit die Zielsetzung des Projekts gefährden. Ohne das erforderliche Update wären nicht nur die Forschungsziele von MissionLab, sondern auch andere dtec-Projekte gefährdet. Um diese Lücke zu schließen, muss die UniBw M ein System beschaffen, das hohe Rechen- und Speicherlasten dauerhaft tragen kann, eine stabile, zertifizierte KI-Plattform bereitstellt und sich nahtlos in die bestehende Forschungsinfrastruktur integrieren lässt.

Official source

Source: TED - Tenders Electronic Daily (Publications Office of the EU)

Always verify details on the official notice.

Last verified against TED - Tenders Electronic Daily (Publications Office of the EU) 3 weeks ago

Market

Software & IT systems in Germany

median award in this market
€415,765
tenders per lot in this market
3.1
More figures

What this market pays

The buyer has too few awards of its own in this sector, so this is the range of the last 613 awards for Software & IT systems in Germany.

Median award
€415,765
Middle half of past awards
€243,493 to €928,512

TED - Tenders Electronic Daily (Publications Office of the EU) publishes no value for this notice, so it has no mark on the scale. The official documents may state one.

Award values as published by the official sources. Use them as a guide, not a rule: scope, lots and duration differ between contracts.

Market context

open now in Software & IT systems
1,865
median open notice
€317,752

Who wins in this market

Many suppliers

Germany · Software & IT systems

different companies won awards here
1,010
of the awards went to the three most frequent winners
11%
buyers awarded contracts here
833

Based on 1,912 awards with a named winner over the last 3 years. Counted by award, not by value, because lot and framework amounts are not comparable. Concentrated means the three most frequent winners took 65% or more of the awards; many suppliers, under 35% with at least ten different winners.

Competition

tenders per lot on average (open and restricted procedures)
3.1
of those lots received a single tender
32%
of the awards went to companies based in another country
5%
of the tenders came from another country
6%

As the official award notices state them, over the last 3 years. Tenders per lot count open and restricted procedures only: in a negotiated procedure or a direct award a single tender is the procedure itself. The country is the one the notice gives for the winner.

The buyer in figures

Universität der Bundeswehr München (UniBw M)

Who wins its contracts

Many suppliers
different companies won its awards
20
of the awards went to the three most frequent winners
19%

Based on 21 awards with a named winner over the last 3 years. Counted by award, not by value, because lot and framework amounts are not comparable. Concentrated means the three most frequent winners took 65% or more of the awards; many suppliers, under 35% with at least ten different winners.

Competition

of the awards went to companies based in another country
10%

As the official award notices state them, over the last 3 years. Tenders per lot count open and restricted procedures only: in a negotiated procedure or a direct award a single tender is the procedure itself. The country is the one the notice gives for the winner.

Universität der Bundeswehr München (UniBw M)

Germany

Similar tenders

Open now

Awarded before

Never miss a tender like this

110 new notices in Software & IT systems in Germany in the last 7 days. Get an email when new ones appear. Free.

Set up my alert

Or follow just its CPV code, Information systems and servers: an email when new tenders arrive under it, no account needed.

Keyboard shortcuts

Read a list without reaching for the mouse. The shortcuts stay quiet while you type in a field.

j
Next result
k
Previous result
Enter
Open the result
n
Next page
p
Previous page
c
Compare this result
s
Shortlist this result
o
Open the result in a new tab
x
Set this result aside
/
Search from any page
Ctrl K
Search from any page
?
Show these shortcuts
Esc
Close

Index

Close
Log in

Language