Data Cities
Data Cities builds and operates AI‑optimized data centers featuring high‑density GPU racks, 400 GbE/CXL networking, NVMe‑over‑Fabric storage, and liquid cooling to deliver high‑throughput, low‑latency compute for large‑scale machine‑learning workloads. Their platform provides automated provisioning, orchestration via SLURM and Kubernetes, and infrastructure‑as‑code tools to enable scalable, power‑efficient training and inference for AI‑first enterprises and research labs.
- Artificial Intelligence
- Hardware
Funding
Founders
Founder details are not available yet.
Product
Problem
AI workloads demand far more compute, bandwidth, and cooling than traditional data center infrastructure, leading many organizations to struggle with performance bottlenecks, high power costs, and limited scalability for large models.
Solution
Data Cities designs and operates AI‑optimized data centers that deliver the full spectrum of artificial‑intelligence compute requirements. Their facilities combine high‑density GPU clusters, 400 GbE networking, NVMe‑over‑Fabric storage, and advanced liquid‑cooling systems to maximize throughput while minimizing power consumption. The architecture is model‑agnostic, allowing customers to run anything from small inference services to massive training jobs without re‑architecting hardware. Integrated management tools such as SLURM, Kubernetes, and infrastructure‑as‑code pipelines provide automated provisioning, resource scheduling, and predictive maintenance, ensuring consistent performance as workloads scale.
Target Audience
Primary customers are AI‑first enterprises, research labs, and cloud providers that require dedicated, high‑performance infrastructure for training and serving large‑scale machine‑learning models.
Features
- High‑density GPU racks with multi‑process service (MPS) support for optimal parallelism
- 400 GbE and CXL interconnects delivering low‑latency, high‑bandwidth communication between accelerators
- NVMe‑over‑Fabric storage tiering for rapid data access and reduced I/O bottlenecks
- Liquid‑cooling infrastructure that lowers energy use and enables seamless scaling of compute density
- Integrated SLURM and Kubernetes orchestration for automated workload placement and scaling
- Infrastructure‑as‑code deployment using Ansible and Terraform for reproducible environment setup
- Predictive maintenance analytics leveraging model‑predictive control to anticipate power and cooling needs