Tenderwize
Retour à la recherche

JARVIS AI Inference Platform – Supply, Installation, and Maintenance

Forschungszentrum Jülich GmbH -Team Infrastruktur-·Jülich, Germany·EUR 23.1M·Expired · Aug 05, 2026·Neg W Call·Supplies
High valueStructured notice only
Titre original de l'avis

Germany – Super computer – JARVIS Eine zentrale Plattform für KI Inferenz • Acquisition of AI inference IT equipment including hardware and mandatory software. • Rack infrastructure, including power distribution units (PDU) and cooling distribution units (CDU) optimized for the offered IT equipment. • Delivery and installation services. • Hardware and software maintenance and support services for five (5) years.

55/100

Score d'opportunité

Needs reviewHigh complexity
SuppliesExpired
Pourquoi ce score ?

Opportunité

55 / 100

Complexité

75 / 100

Facteurs de risque

No procurement documents (e.g., technical concept, evaluation matrix) were analyzed; critical details missing.

Submission deadline is 2026-08-05; limited time to prepare a technically complex bid.

No procurement documents available

Analyse approfondie du portefeuille

Analyse approfondie de l'adéquation entreprise

Utilise:

  • Description
  • Industries & Services
  • Capabilities
  • Market & Experience
  • Certifications

Aperçu

Moderate opportunity

Faits clés

  • EUR 23.1M estimated value
  • Supplies contract
  • 1 lot
  • Duration 5 YEAR
  • Supply of AI inference IT equipment (hardware and mandatory software)
  • Provision of rack infrastructure including PDUs and CDUs optimized for the offered equipment
Afficher le résumé complet

This tender by Forschungszentrum Jülich GmbH seeks to acquire a heterogeneous AI inference platform (JARVIS) integrated with the JUPITER exascale supercomputer. The procurement includes AI inference IT equipment (hardware and software), rack infrastructure (PDUs, CDUs), delivery/installation, and five years of maintenance. The single lot is valued at €23.1 million, EU-funded, and uses a negotiated procedure with a call for competition. The deadline for submissions is 2026-08-05 10:00 CEST. No procurement documents beyond the TED notice were analyzed, so detailed eligibility and award criteria are unavailable.

Risques

No procurement documents (e.g., technical concept, evaluation matrix) were analyzed; critical details missing.

Submission deadline is 2026-08-05; limited time to prepare a technically complex bid.

No procurement documents available

L'analyse peut être incomplète

Seule une partie des documents de consultation a été analysée. Des exigences, certificats ou documents de soumission supplémentaires peuvent exister dans les documents restants.

Exigences clés

Technical

  • Heterogeneous system supporting AI inference workloads
  • High-speed, low-latency interconnect for large compute pod

Administrative

  • Integration with existing JUPITER infrastructure and cloud platform

Les exigences peuvent être incomplètes.

Critères d'attribution

Lot 1

Quality-based award criteria with 0-10 point scale per criterion. Detailed evaluation matrix not available.

Acheteur

Localisation

Jülich, DEU

Profil acheteur

Ouvrir le profil

Identifiant

Berichtseinheit-ID 00013306

Activité

Education

Lots (1)

LOT-0001

JARVIS AI Inference Platform – Supply, Installation, and Maintenance

Acquisition of a heterogeneous system for AI inference, including high-speed compute pods, confidential computing resources, Ethernet interconnect, and European technology preference. Scope covers hardware, software, rack infrastructure, delivery, installation, and 5-year maintenance. Integration with existing JUPITER infrastructure.

SuppliesEU fundedElectronic submission

EUR 23.1M

Valeur estimée

Localisation

Jülich, Germany

Durée

Feb 2027

Catégorie

Super computer

Critères d'attribution

Quality

Date limite

Expired Aug 05, 2026

Détails de la procédure

Date de publication
30 Jun 2026
Langues
German

Métadonnées de référence

Base juridique
32014L0024

Identifiants de référence

ID de l'appel d'offres
450843-2026

Documents (1)

Appels d'offres similaires