100G DAC Cables
100G QSFP28 passive DAC cables, 0.5 m to 3 m, coded for Cisco Nexus, Arista, Juniper QFX, Dell PowerSwitch and NVIDIA/Mellanox Spectrum switches. Also here: EDR InfiniBand versions for ConnectX-5 and Quantum EDR, QSFP28-to-4×SFP28 breakouts that split one 100G port into four 25G links, and 200G-to-2×100G breakouts. New and refurbished, lifetime warranty, shipping from Knoxville, Tennessee.
Other speeds: 25G · 200G · 400G · 800G · All DAC cables
New to DAC? Read our DAC Cables Guide →
From the blog
Why engineers and teams are moving LLM inference and fine-tuning onto DGX Spark and GB10 systems: bigger models than one...
EPYC vs Xeon for GPU Inference Servers: Dell R7525 vs R740xd
Does the CPU matter for GPU inference? Dell R7525 (EPYC) vs R740xd (Xeon) compared on PCIe, memory bandwidth and GPU...
On-Prem AI Inference Fabric: Do You Need a Data Center Switch?
When on-prem AI inference needs a data center switch, when a DAC or ring is enough, and what lossless RoCE...
Why Teams Are Bringing LLMs In-House on DGX Spark and GB10: Privacy, Fine-Tuning and Control
Why engineers and teams are moving LLM inference and fine-tuning onto DGX Spark and GB10 systems: bigger models than one unit can hold, private data that never leaves the building, and a desk-side box that matches NVIDIA's production stack.
EPYC vs Xeon for GPU Inference Servers: Dell R7525 vs R740xd
Does the CPU matter for GPU inference? Dell R7525 (EPYC) vs R740xd (Xeon) compared on PCIe, memory bandwidth and GPU count.
On-Prem AI Inference Fabric: Do You Need a Data Center Switch?
When on-prem AI inference needs a data center switch, when a DAC or ring is enough, and what lossless RoCE Ethernet requires.
Testimonials
Got to me in 2 days, looks as professional as CDW.