GREENPUTE · MAINTENANCE

High-end GPU server maintenance

Factory-grade and third-party maintenance for NVIDIA, AMD, SMC, Dell, Asus, GBT and other mainstream brands. Hardware repair, firmware upgrades and performance tuning, with round-the-clock cover and a 2-hour average response.

Maintenance is the core business. Board-level repair, high-density rooms and liquid cooling exist to keep compute deliverable.

01 / SLA

7×24

Maintenance cover

02 / SLA

2 hours

Average response

03 / SLA

99.99%

Operational stability

Brands under maintenance

  • NVIDIA
  • AMD
  • Supermicro
  • Dell
  • ASUS
  • GIGABYTE
  • HP
  • ASRock

MAINTENANCE

How maintenance is delivered

National centers take the daily load. Faults that can be judged on site stay on site. Parts that must be replaced go to level-3 repair.

  1. 01

    National centers

    Maintenance centers in intelligent-computing campuses, with 7×24 cover and tiered repair.

  2. 02

    On-site level 2

    Engineers diagnose on site. Faults that can be cleared in the room do not wait overnight.

  3. 03

    Depot level 3

    Boards, GPUs and power modules move to level-3 repair, cutting spare-parts and downtime cost.

  4. 04

    Spares and SLA

    A thousand-square-meter spare pool plus direct international sourcing. Response follows the SLA.

RESPONSE

Someone picks up when a machine fails

Specialists stay online 24 hours. Dispatch reaches repair stations in more than one city.

24h

Specialists online

0 h

Average response

0.00%

Stability

0+

Custom-ops clients

ALSO

Three things beside maintenance

Repair brings a single machine back. Rooms and liquid cooling keep a whole hall of compute running.

Board-level repair

Diagnose and repair motherboards, GPU cores and power modules. Repair centers sit at large compute nodes, with parts replacement at level 3.

High-density rooms

Consulting, design and handover for data centers, especially high-density compute halls.

Liquid cooling

Design, integration and operations for immersion and cold-plate systems, aimed at high-power heat and a lower PUE.

We are not only the people who maintain the equipment. We are the partner who keeps compute steady and energy use in check.