# Silicon Network

> Dedicated NVIDIA GPU compute on long-term contract. Operated by Berkeley Compute Inc.

## Canonical facts

- Website: https://silicon.net
- Contact: https://silicon.net/contact
- Tagline: Build AI on dedicated compute
- Business model: Silicon owns GPU hardware; customers reserve dedicated compute under multi-year contracts
- Uptime SLA: 99.9%
- Datacenter tier: Tier 3 certified datacenters
- Regions: Oregon, Texas, Arizona (United States)

## Clusters coming online

- Oregon: 72 × NVIDIA H100 SXM — $2.00/gpu/hour, 24-month term, 25% prepayment, 60-day go-live
- Texas: 128 × NVIDIA B200 — $3.70/gpu/hour, 36-month term, 20% prepayment, 90-day go-live
- Arizona: 64 × NVIDIA B300 — $4.00/gpu/hour, 36-month term, 20% prepayment, 90-day go-live

## Frequently asked questions

### What is Silicon Network?
Silicon Network provides dedicated NVIDIA GPU compute on multi-year contracts in Tier 3 datacenters.

### What is dedicated compute?
Committed GPU capacity under contract, with predictable pricing and SLAs, instead of spot or shared cloud instances.

### How do I reserve capacity?
Use https://silicon.net/contact or the Reserve Capacity call-to-action on the homepage.

## Key pages

- Home: https://silicon.net/
- Contact: https://silicon.net/contact
- Blog: https://silicon.net/blog
- Privacy: https://silicon.net/privacy-policy
- Terms: https://silicon.net/terms-of-service