5,000-GPU cluster network fabric
Company: Applied Digital
The claim, verbatim
An Applied Digital 5,000 H100 GPU cluster uses 250 km of fiber and more than 15,000 InfiniBand cables to interconnect all GPUs.
Source (primary)
applied-digital 8-K exhibit (technicaloverview1012202.htm) (SEC EDGAR (8-K Ex), press_release)
View cached copy (2026-11-18)Live source ↗
Quote: “An APLD 5,000 H100 GPU cluster consists of 250km of fiber interconnecting all GPUs for a total of 15k+ InfiniBand cables”
How we checked this
Checked on November 20, 2026. The cited source supports every part of this claim.
The presentation states as a fact that an APLD 5,000 H100 GPU cluster has 250km of fiber interconnecting all GPUs, totaling 15k+ InfiniBand cables.
Confirmed in the source:
- An Applied Digital 5,000 H100 GPU cluster uses 250 km of fiber to interconnect all GPUs
- Such a cluster uses more than 15,000 InfiniBand cables
What we did: Read our cached copy of SEC EDGAR (8-K Ex) (https://www.sec.gov/Archives/edgar/data/1144879/000114487923000216/technicaloverview1012202.htm) in full (11,418 characters, retrieved November 18, 2026) and checked each assertion in the claim against it.
Additional evidence
confirms applied-digital 8-K exhibit (technicaloverview1012202.htm)
Quote: “An APLD 5,000 H100 GPU cluster consists of 250km of fiber interconnecting all GPUs for a total of 15k+ InfiniBand cables”