Plutus for Treasury and Liquidity: Sovereign Finance Intelligence Behind the Firewall

By Micky Irons, founder and CEO of Mickai

The treasury problem nobody wants to put in the cloud

Ask any bank treasurer where their liquidity model runs and you will hear the same hesitation. The intelligence that decides intraday funding, capital allocation and the liquidity coverage ratio sits on the most sensitive data a regulated institution holds: counterparty exposures, deposit behaviour, payment flows, balance-sheet positions. That data is exactly what the Prudential Regulation Authority expects you to govern, evidence and keep inside controlled boundaries under SS2/21. So the modern AI that could transform treasury work has, until now, been quietly fenced out of the one place it would be most valuable.

Plutus is built to close that gap. It is the finance Studio inside Mickai, the sovereign AI operating system, and it runs cashflow forecasting, liquidity intelligence and capital planning entirely on-premises, inside the institution's own walls, air-gapped where the risk profile demands it. The CFO office gets a finance brain with real analytical horsepower, and the regulator gets something it can actually accept: a system where no special-category or material non-public data ever leaves the building. This is built and live today, deployed inside the institution's perimeter rather than promised for a later release.

Why on-prem is the unlock, not a compromise

The instinct in some quarters is to treat sovereign deployment as the cautious, slower option. The opposite is true. The reason treasury AI has stalled is that public-cloud models force a choice between capability and control, and regulated finance cannot give up control. Roughly 0.85 million UK businesses, about 15 percent of the market, sit under rules that bar them from putting this class of data into public-cloud AI. Across the EU the figure is around 5 million. PRA SS2/21, UK GDPR special-category handling, the EU AI Act high-risk regime and the CLOUD Act jurisdictional reach all push the same way.

Mickai treats that constraint as the design centre rather than an afterthought. Plutus runs where the data already lives. Models execute inside the institution's perimeter, on hardware the institution owns. Nothing is shipped to a third-party endpoint for inference. That is what makes the difference between a pilot a risk committee tolerates and a system a board is willing to run in production. The sovereign AI market is moving the same direction, from around USD 40 billion in 2025 toward an estimated USD 148 billion by 2032, because the institutions with the most valuable data are the ones least able to use the public-cloud version.

What Plutus actually does for the CFO office

Plutus is not a chat box bolted onto a spreadsheet. It is a finance intelligence layer with three working surfaces.

Cashflow forecasting takes payment flows, receivables, deposit behaviour and scheduled obligations and builds forward views the treasury team can interrogate, stress and revise, with the model reasoning over the institution's own history rather than a generic prior.

Liquidity intelligence keeps the regulatory ratios in view as live, explainable positions. Coverage ratios, funding concentrations and intraday liquidity are surfaced with the drivers behind them, so the team sees not just the number but why it moved and what would move it back.

Capital planning ties it together, letting the office model allocation, run scenarios against the balance sheet and pressure-test plans before they reach a committee. Because Plutus sits inside Mickai, it works alongside the other Studios. Prometheus contributes forecasting depth, Nomos carries the compliance logic, Nemesis watches for the fraud and AML signals that distort liquidity assumptions, and Aletheia handles the audit trail. The treasury team is not buying ten point tools that do not talk to each other. They are running one operating system with finance as a first-class function.

The audit record regulators were waiting for

The part that turns this from interesting to acceptable is the Operational Audit Record. Every action Plutus takes, every model run, every input, every output that informs a treasury decision, is written to a tamper-evident, post-quantum-signed audit record. When the PRA asks how a liquidity figure was produced, the answer is not a verbal reconstruction months later. It is a signed, immutable record that shows the inputs, the model state and the reasoning, end to end.

This is the difference between AI that a regulated treasury can defend and AI that creates a governance liability. SS2/21 is, at heart, about being able to evidence model risk management. An auditable-by-design system answers that requirement structurally rather than with a manual paper chase after the fact. Sovereignty keeps the data in; the audit record proves what the system did with it.

A moat built on filings, not promises

Behind Plutus sits an intellectual property estate that explains why this is hard to copy. Mickai LTD holds 104 filed UK patent applications carrying roughly 2,340 claims, with myself as named inventor. These are filed, not granted, which is the honest framing: what they establish is priority and a prior-art position across sovereign on-prem AI, the audit record architecture and the Studio model. That estate is broad enough that 196 companies map to 311 patent-company pairs as potential licensees, including the largest names in cloud and silicon such as Microsoft, AWS, NVIDIA, Google, Adobe and IBM. That is potential-licensee sizing rather than booked revenue, but it tells you where the technical centre of gravity is sitting.

As a third-party momentum signal, in June 2026 I was ranked number 4 on Crunchbase, with Mickai placing in the top 1 to 2 percent of companies globally. We are a UK company with Birmingham manufacturing secured, building to scale.

Where this fits in the bigger picture

Mickai is an ally to the broader AI industry, not a challenger trying to out-shout it. The hyperscalers serve the open, public-cloud world brilliantly. What they cannot easily serve is the regulated institution that is legally barred from using that world for its most sensitive workloads. Plutus, and Mickai around it, is the layer that lets those institutions adopt frontier-grade finance intelligence without breaking the rules they are bound by.

That dual-buyer thesis, serving the regulated enterprise directly while standing in a category a hyperscaler would strategically want to own, is what underwrites the value here. The Year 5 revenue path runs to billions at high gross margin, anchored by the IP estate rather than by hype. The technical centre of gravity in sovereign on-prem finance AI is sitting on filed priority, a live product and a manufacturing base, which is a rare combination in this category.

A narrow window for selected partners

We are opening early collaboration to a small number of treasury and finance teams who want to run Plutus against their own balance sheet, inside their own walls, before the category fills in around them. This is selective by design, not a general call. If you run a treasury function under PRA SS2/21 or an equivalent regime and you have been waiting for finance AI that the regulator will actually accept, this is the moment to talk.

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

FAQ

See below.

Frequently asked questions

Can Plutus run without sending treasury data to the cloud?

Yes. Plutus runs entirely on-premises inside the institution's own perimeter, on hardware the institution owns, and air-gapped where the risk profile demands it. No special-category or material non-public data is shipped to a third-party endpoint for inference. The data stays where it already lives.

How does Plutus help with PRA SS2/21 compliance?

SS2/21 is about being able to govern and evidence model risk management. Plutus keeps the sensitive data inside controlled boundaries and writes every model run, input and output to a tamper-evident, post-quantum-signed Operational Audit Record, so a liquidity figure can be reconstructed as a signed, immutable trail rather than a verbal account after the fact.

What does Plutus actually do for a treasury team?

Three things: cashflow forecasting over the institution's own history, liquidity intelligence that keeps regulatory ratios live and explainable, and capital planning with scenario and stress testing against the balance sheet. Because it sits inside Mickai, it works alongside the other Studios for forecasting, compliance, fraud and AML, and audit.

Is Plutus a live product or a roadmap item?

It is built and live, deployed inside the institution's perimeter today. Mickai is the sovereign AI operating system around it, with Plutus as the finance Studio, and the company is building to scale from a UK base with Birmingham manufacturing secured.

What protects Mickai's position in this category?

Mickai LTD holds 104 filed UK patent applications carrying roughly 2,340 claims, establishing priority and a prior-art position across sovereign on-prem AI, the audit record architecture and the Studio model. That estate maps to 196 companies as potential licensees across cloud and silicon, which indicates where the technical centre of gravity sits.