Insights from the AI Lab
Sponsored by Dell Technologies
Fresh insights from our datacenter in Colorado. Bringing real-world performance and analysis with hands-on testing of enterprise technologies.
Network Topology Optimization for AI Workloads
From bottleneck to breakthrough. Dell PowerEdge XE9680 with Broadcom networking delivers congestion-aware fabrics that improve efficiency, streamline training, and turn network optimization into a competitive advantage.
- Brian Martin
AI Networking
Server Platforms
Scaling AI with Dell PowerEdge XE9680
Choice built in. Dell PowerEdge XE9680 with Broadcom networking empowers enterprises to scale AI their way and balance performance, cost, and ecosystem flexibility on one unified platform.
- Brian Martin
GPU Acceleration
Server Platforms
Dell AMD Instinct Series GPU Cluster with Dell Networking
Accelerate innovation. With AMD Instinct GPUs and Broadcom Ethernet, the Dell PowerEdge XE9680 powers faster model training, efficient scaling, and simplified operations for AI at scale.
- Brian Martin
AI Networking
GPU Acceleration
Server Platforms
Dell PowerEdge XE9680 H200 Cluster with Dell 400GbE Networking
Engineered for efficiency and speed. The Dell PowerEdge XE9680 with NVIDIA H200 GPUs and Broadcom Ethernet achieves near perfect scaling and low latency throughput for AI training and inference.
- Brian Martin
AI Networking
GPU Acceleration
Server Platforms
AI On-Premises: A Look at OpenAI GPT-OSS-120B
Bring AI in-house without compromise. OpenAI’s OSS-120B model, tested on Dell PowerEdge with NVIDIA and AMD GPUs, proves enterprises can achieve cloud-class performance on-prem while keeping full control of their data.
- Brian Martin
Emerging Tech
GPU Acceleration
Latest Insights
About the Lab
Introducing our AI Performance and PoC lab in Colorado:
- Built for benchmarking, iteration, and real-world results
- 3MW+ capacity with air-cooled systems today, liquid-cooled tomorrow
- Supports LLM training and inference, RAG pipelines, and computer vision
- Ready for next-gen accelerators with 400/800G fabrics, high-memory nodes, and scalable NVMe/object storage
- Engineers get secure remote access, reproducible testbeds, and deep observability
- Optimized for altitude efficiency and deployment flexibility
- Mission: Acceleration innovation to real-world AI impact
