Built in Newcastle. Running in your city.
Newcastle Rising builds custom AI-powered applications for professional services firms, running on dedicated hardware co-located in the client's city. The AI is frontier-grade. The data never leaves the client's jurisdiction.
The managed service, plainly stated.
Most businesses using AI today route their data to servers in the United States — through ChatGPT, Microsoft Copilot, or a cloud-hosted AI SaaS product. For many professional services firms, that's a problem: client confidentiality obligations, patient data regulations, or simply the reasonable preference to control where sensitive information lives.
We solve that by building the full stack ourselves: we procure and co-locate a dedicated GPU server at a professional datacenter in the client's city, deploy a frontier open-source AI model on it, build custom applications tailored to the client's specific workflows, and manage the whole system ongoing. The client gets frontier AI capability in purpose-built tools. Their data never moves.
The result is something no existing product provides: frontier model quality, running locally in the client's city, on hardware they control, customised to their specific documents and workflows, maintained by a specialist, at a predictable monthly cost, with zero dependency on any US technology company.
A genuinely uncontested market.
Every other AI company in most cities is reselling OpenAI, Anthropic, or Microsoft. Their entire offering is an API key and some configuration. The client's data goes to servers in the United States. The vendor has no technical moat.
The sovereign AI managed service position sits in a gap that other categories cannot fill:
- Too small for large cloud vendors — they need cloud dependency
- Too technically complex for local IT support companies
- Too privacy-sensitive for generic cloud AI SaaS startups
- Completely unaddressed in most mid-sized cities
The window will narrow as the technology matures. Early clients get switching cost and workflow depth that compounds over time — and reference clients in a professional network are worth more than any marketing spend.
The broader project.
Newcastle Rising is a deliberate effort to build digital infrastructure in regional Australia that gets used by the people who live here, rather than serving as a node in someone else's network. The sovereign AI work is the strand that matters most commercially — it's what funds everything else and demonstrates that local technical capability can operate at a frontier level.
Other strands — housing, civic data, transport — sit under the broader Newcastle Rising umbrella at newcastlerising.com.au. The compute work stands on its own commercially.
Who you're working with.
Matt Cook — Founder
Software engineer, Hunter resident, founder of Newcastle Rising. Background in shipping production AI and web systems across multiple ventures. Operates from Newcastle; reachable directly.
The team is deliberately small: senior engineers, all Australia-based, all hands-on. The strategy depends on cycle time, and cycle time depends on staying small and capable. You get the people who do the work, not an account manager standing between you and them.
Operating philosophy
Honest about what's built and what's not. Cost-plus pricing rather than capture-pricing. Reinvest surplus into capability rather than distribute to remote shareholders. Move fast on technology; move slow on commitments.
If something we publish here turns out to be wrong, we change it and say what changed. No vanity metrics, no fabricated client logos, no theatre.
What's next.
Now — building first clients
Sovereign AI managed service for professional services firms. Custom application builds for law, accountancy, healthcare, and other sensitive-data verticals. Each client is a reference; each reference opens the next conversation.
6–12 months — vertical depth
Deep specialisation in two or three verticals where the application library is richest, fine-tuning the model on real client data that has accumulated, and establishing channel partnerships with IT managed service providers who already have the client relationships.
12–24 months — geographic expansion
Replicating the model in additional cities and regions. Each new city is largely self-funding once the first few clients come on — the hardware economics are predictable and each server increment is small.