Prometheus: Air-Gapped Demand and Risk Forecasting for Regulated Operators

The forecasting problem nobody wants to admit

Every regulated operator I speak to has the same quiet anxiety. They want better forecasting. They want to model demand swings, supply shocks, counterparty exposure and capital allocation under dozens of scenarios at once. And they cannot do it the way the market keeps telling them to, because doing it means handing their most sensitive operating data to a public-cloud model sitting outside their walls and outside their control.

For a grid operator, that data is real-time load and generation profiles. For a manufacturer, it is order books, supplier terms and inventory positions. For a lender or insurer, it is exposure, pricing and counterparty health. This is the information that defines competitive position and, in many cases, national resilience. Sending it to a third party to get a forecast back is not a trade-off most boards will sign. Under PRA SS2/21 on model risk, UK GDPR special-category handling, the NIS Regulations and the EU AI Act high-risk regime, it is often not even legal.

So the forecasting stays primitive. Spreadsheets, a handful of legacy models, gut feel. The capability gap is real and it is expensive.

Prometheus closes that gap without asking anyone to compromise on sovereignty.

What Prometheus is

Prometheus is the forecasting Studio inside Mickai, the sovereign AI operating system. Mickai is AI that regulated businesses own and run inside their own walls, on-premises and air-gapped, with every action written to a tamper-evident, post-quantum-signed audit record we call the OAR. This is built and live, not a concept on a roadmap.

Prometheus runs scenario forecasting for demand, supply and capital planning entirely inside that boundary. The data never leaves. The model never calls home. There is no public-cloud endpoint in the path, no vendor telemetry, no shadow copy of your operating data sitting in someone else's region waiting for a US CLOUD Act request or a breach disclosure.

Named for the Titan who saw what was coming, Prometheus does the one thing forecasting is supposed to do and does it where it is safe to do it.

How it works inside the wall

Prometheus ingests your historical and live operating data from inside your own network. It builds probabilistic forecasts across the variables that matter to your sector: energy load and generation, raw-material and component supply, order demand, working capital, counterparty and market exposure. It runs these as scenarios rather than single point estimates, so you see the distribution of outcomes and the tails, not a falsely confident line.

Three things make it usable for a regulated operator rather than just interesting.

First, it is air-gapped by design. Prometheus is not a cloud service with an on-prem wrapper. It is part of an operating system that assumes the network boundary is sealed and works anyway. That is the whole point of Mickai.

Second, every forecast run, every input set, every assumption and every output is written to the OAR. When a regulator, an internal model-risk function or an auditor asks how a number was produced, you have a tamper-evident, post-quantum-signed record of exactly that. SS2/21 expects you to evidence model governance. Prometheus produces that evidence as a by-product of running.

Third, it sits alongside the rest of the Studio modules. A demand forecast from Prometheus can feed Plutus for FP&A, Tyche for underwriting, or Nemesis for fraud and AML signals, all inside the same sealed environment with one shared audit trail. You are not stitching together a forecast from one vendor with planning from another and hoping the data lineage holds.

Why sovereign forecasting is a market, not a niche

The firms that legally cannot send data to public-cloud AI are not a rounding error. We size the wedge at roughly 850,000 UK businesses, about 15 percent of the total, and around 5 million across the EU. The drivers are not going away: PRA SS2/21, UK GDPR special-category rules, the NHS DSP Toolkit, the EU AI Act high-risk classification, ITAR and EAR, the NIS Regulations and the extraterritorial reach of the US CLOUD Act. The sovereign AI market was around 40 billion US dollars in 2025 and is projected to reach 148 billion by 2032.

Forecasting is one of the clearest expressions of that demand. It needs the most sensitive data to be useful, which is exactly why it has been hardest to modernise under the public-cloud model. Prometheus is built for the operators who were told they had to choose between capability and control, and refused.

As a dated, third-party momentum signal: in June 2026, Micky Irons was ranked number four on Crunchbase's CB Rank for people, verified live, with the Mickai company profile in the top one to two percent globally. I share that as an outside reading of momentum at a point in time, not as a permanent claim. The thesis it reflects is simple. Serious regulated demand is heading toward sovereign AI.

Mickai is an ally, not a replacement

I want to be precise about positioning, because it shapes how Prometheus fits your stack. Mickai is not here to replace the frontier labs or tell you to rip out the public cloud. For workloads that are not sensitive, those tools are excellent and you should keep using them. Mickai is the place you run the workloads you legally and commercially cannot put anywhere else. That is the dual-buyer thesis. Public cloud for the open work, Mickai for the sovereign work. Prometheus is the forecasting expression of that boundary.

Underneath it sits real intellectual property. Mickai LTD holds 104 filed UK patent applications covering roughly 2,340 claims, with myself as inventor. These are filed, not granted, and I am clear about that distinction. They establish priority and a prior-art moat around sovereign, audited, air-gapped AI operation. It is the architecture, not a marketing line.

The pre-seed window

Mickai is built and live, and we are building to scale. UK company, Birmingham manufacturing secured, Studios including Prometheus running inside the operating system today.

As we scale, a pre-seed window is open to a small number of selected partners. This is an opportunity to get involved early in sovereign AI infrastructure for regulated operators, not a search for rescue. If you run forecasting for an energy, supply-chain, financial or critical-infrastructure operation and you have been waiting for a way to do it without surrendering your data, or if you want to be part of the build as a partner, I would like to talk.

Reach me directly at micky@mickai.co.uk.

Micky Irons, founder and CEO of Mickai.

Frequently asked questions

Does Prometheus send any of our operating data to the cloud?

No. Prometheus runs entirely inside your own walls, on-premises and air-gapped, as part of the Mickai sovereign AI operating system. Your historical and live operating data stays inside your network. There is no public-cloud endpoint in the path, no vendor telemetry and no external copy of your data.

How does Prometheus help with model-risk and regulatory evidence?

Every forecast run, including its inputs, assumptions and outputs, is written to the OAR, a tamper-evident, post-quantum-signed audit record. When a regulator or an internal model-risk function asks how a number was produced, you can evidence it directly. This maps to expectations under PRA SS2/21 on model governance, and the record is produced as a by-product of running rather than reconstructed after the fact.

Is Prometheus built and available now, or on a roadmap?

It is built and live. Prometheus runs today inside the Mickai operating system as one of the Greek-named Studio modules, alongside Plutus, Tyche, Nemesis and others. Mickai is a UK company with Birmingham manufacturing secured, building to scale.

Does Mickai replace tools like the public-cloud frontier labs?

No. Mickai is an ally, not a replacement. For non-sensitive workloads, public-cloud AI is excellent and you should keep using it. Mickai is where you run the workloads you legally and commercially cannot put anywhere else. That is the dual-buyer thesis: public cloud for the open work, Mickai for the sovereign work.

How can a forecasting operator or partner get involved?

A pre-seed window is open to a small number of selected partners as Mickai scales. If you run forecasting for an energy, supply-chain, financial or critical-infrastructure operation, or want to partner on the build, contact Micky Irons directly at micky@mickai.co.uk.