Spark AI Foundry

About

Building the infrastructure backbone for the AI era. A continental grid of high-performance compute sites interconnected by a TeraB it network core.

Our Mission

A Grid Network Built for AI

Spark AI Foundry is developing a grid network of 20+ AI Foundry sites across North America dedicated to high-performance AI compute and large-scale data storage, interconnected with a TeraB it network core.

These sites are being developed as part of Spark AI's Neocloud architecture, delivering scalable, power-dense infrastructure that supports the full range of modern AI workloads: from frontier model training to real-time inference, enterprise AI applications, and research computing.

Each Foundry site is engineered for the demands of AI infrastructure: high power density, redundant fiber interconnect, and operational excellence, forming a distributed compute fabric across the continent.

20+
AI Foundry Sites
TeraBit
Network Core
North America
Coverage
Neocloud
Architecture
TeraB it Network Core

All Foundry sites are interconnected via a high-capacity, low-latency TeraB it backbone, enabling workload distribution and data mobility across the entire grid.

Neocloud Architecture

Infrastructure for Every AI Workload

Spark AI's Neocloud architecture supports the full range of AI and advanced compute workloads, from large frontier model training runs to real-time enterprise applications.

Large Language Models

Power density and low-latency interconnect for training and serving frontier LLMs at scale.

Generative AI

High-throughput GPU clusters for image, video, audio, and multimodal generative workloads.

Real-Time Inference

Low-latency edge-adjacent nodes for sub-millisecond AI inference in production applications.

Research Computing

HPC environments supporting academic institutions, national labs, and frontier AI research.

Enterprise AI Supply Chain

AI-driven supply chain optimization, demand forecasting, and logistics tools for enterprise customers.

Pharma & Digital Twins

Compute-intensive simulation environments for pharmaceutical R&D, molecular modeling, and digital twin work.

Active Developments

Current Foundry Sites

Loveland, CO

221,366 sq ft | 38.467 acres | 16,000A power

Former Woodward high-tech manufacturing campus being converted into an AI data center and advanced technology hub.

View Location

Olathe, KS

~160 acres | Up to 500 MW | BNSF rail access

Greenfield campus in Johnson County at the geographic center of the US, with large power capacity and central network positioning.

View Location

Lansing Township, MI

105.25 acres | Abutting substation | Active rail

Former GM industrial land in active development with Deep Green for high-density, sustainable AI colocation deploying in weeks at up to 150kW per rack.

View Location

Leadership

Team

JP

Jack Pressman

Chief Technology Officer

Jack leads technology strategy and infrastructure development at Spark Data Center Services, LLC, overseeing the design and execution of Spark AI Foundry's continental Neocloud architecture.

Get Involved

Partner with Spark AI Foundry

We're building the AI infrastructure backbone for North America. Whether you're an investor, technology partner, or enterprise customer, we'd like to hear from you.