Own It, Do Not Rent It: Sovereign Intelligence for the Regulated Enterprise

The distinction that decides everything: ownership

Mickai is a sovereign intelligence operating system that regulated businesses own and run inside their own walls. That single sentence is the whole argument. Every other AI system a regulated firm can buy today arrives as a tenancy. The model sits on someone else's infrastructure, the data travels to a jurisdiction the buyer does not control, and the terms of service can change under the buyer's feet. We built the alternative, and it is live now, not a roadmap item and not a pilot waiting for funding to begin.

We run entirely on the customer's own hardware, on premises and air gapped, with zero data egress and no public cloud round trip. Nothing leaves the building. That is not a configuration option we bolt on for nervous clients. It is the architecture. When a bank, an insurer, a hospital or a defence contractor deploys us, the intelligence lives where the regulated data already lives, and it never has to leave to be useful.

Why renting fails the regulated buyer

There is a large class of organisation that cannot lawfully send its data to public cloud AI. This is not a matter of preference or caution. It is a matter of law and enforceable regulation. We count around 0.85 million UK businesses, about 15 percent of the total, that fall into this position, and roughly 5 million across the EU. These are not fringe operators. They are the institutions that hold the most sensitive data in the economy.

The drivers are specific and well documented. The Prudential Regulation Authority sets model risk expectations under SS1/23. UK GDPR governs special category data. The NHS Data Security and Protection Toolkit binds anyone touching patient records. The EU AI Act places high-risk classification on exactly the systems these firms want to build. ITAR and EAR govern defence and dual-use technology. The NIS Regulations cover critical infrastructure. And the US CLOUD Act means that data held by an American hyperscaler can be reached by a foreign government regardless of where the servers physically sit. For a firm bound by any one of these, a public cloud AI tenancy is not a bargain to be negotiated. It is a door that is legally shut.

We built Mickai for the organisations standing on the wrong side of that door. When the intelligence runs on the customer's own hardware, air gapped, with no egress, the legal objection does not need to be managed. It simply does not arise.

What we actually run

Underneath the operating system we run about fifty specialist models, twenty-five domain and twenty-five operational, with cross-model routing under a deterministic arbiter. That last point matters more than it may first appear. Determinism means outputs are reproducible. Ask the same question of the same state twice and you get the same answer, which is precisely what a regulator, an auditor or a court expects and precisely what a probabilistic cloud endpoint cannot promise. We refer only to our own models, the sovereign models, trained and held for this purpose.

The capabilities are organised into studios, each named from Greek myth and each doing serious, narrow work. Nemesis handles fraud and anti money laundering. Plutus covers finance and FP&A. Tyche does underwriting. Prometheus does forecasting. Iris runs customer service. Nomos handles compliance, Astraea covers legal, and Panacea covers clinical work. Pythia delivers business intelligence, Aletheia handles audit, and Vinis provides voice. The Agentic Marketing Team runs marketing operations, Trust Agent holds the perimeter, and we offer OAR-as-a-Service for organisations that want the audit substrate on its own.

Proof you can verify without trusting us

Ownership is only half the promise. The other half is proof. Every consequential action inside Mickai is written to the Open Audit Record. Each entry is signed under post-quantum cryptography, FIPS 204 ML-DSA-65 with ML-KEM-768, and hash-chained into a tamper-evident, append-only ledger. Anyone can verify that ledger offline, for decades, without trusting the vendor. Read that clause again, because it inverts the usual relationship. We are not asking a regulated buyer to take our word for what the system did. We are handing them a record they can check themselves, cryptographically, long after any commercial relationship has ended.

For deployments that span multiple sites or fielded units, Pantheon extends the same guarantee across nodes. It is a post-quantum Layer 1, currently on testnet, that provides multi-node attestation with no central server. Each unit can attest to the others without a coordinating authority in the middle, which is the property you need when the whole point is that no single controller sits above the estate.

The intellectual property position

We have filed 104 UK patent applications, comprising roughly 2,340 claims across 13 invention families, owned by Mickai LTD, with named inventor Mickarle Sean Junior Wagstaff-Irons. These are filed, not granted. We are precise about that distinction because it matters. Filing establishes priority and builds a prior-art moat. It fixes the date and stakes the ground. We do not claim more than the filing gives us, and we do not need to.

Two buyers, one stack

There are two ways to reach the regulated market, and we pursue both. The first is direct: we sell sovereign AI to the firms the public cloud cannot lawfully serve. The second is by licence: we make the patented stack available to the platforms that want to reach those same firms and currently cannot. A platform that adds a sovereign layer to its offering reaches, in a single step, the regulated market it is otherwise legally barred from serving. That is a plain statement of where the value sits.

Our internal analysis maps 196 companies and 311 patent-company pairs as potential licensees, and the names in that set include Microsoft, AWS, NVIDIA, Google, Adobe and IBM. We are careful about what that is and is not. It is potential-licensee sizing, a map of where our patent families touch existing product lines. It is not a signed book of business and it is not an infringement claim against anyone. We position ourselves as an ally to the AI majors, not as an adversary. The public cloud does extraordinary things for the markets it can serve. We serve the markets it cannot, and we make it possible for the majors to follow.

The market we are building into

The sovereign AI market sits at roughly USD 40 billion in 2025 and is projected to reach about USD 148 billion by 2032. That trajectory is not driven by fashion. It is driven by the regulatory drivers above, each of which tightens rather than loosens over time. As the EU AI Act phases in, as model-risk supervision matures, and as data-sovereignty enforcement grows teeth, the population of organisations that must own rather than rent their intelligence expands. We are built for that population, and we are already deployable to it.

The company

Mickai LTD is a UK company, Companies House number 17166618, with Birmingham manufacturing secured. Micky Irons is founder and CEO. We build the operating system, the models, the audit substrate and the attestation layer as one coherent stack, because sovereignty that is only partial is not sovereignty at all.

The message to a regulated buyer is simple. You should own the intelligence that touches your most sensitive data, run it inside your own walls, and hold a record of what it did that you can verify yourself, without trusting anyone. That is what we built. You do not rent it. You own it.

Our pre-seed round is opening soon, and we welcome inquiries from interested partners by email at micky@mickai.co.uk or on LinkedIn.

Does Mickai really run with no connection to the public cloud?

Yes. We run entirely on the customer's own hardware, on premises and air gapped, with zero data egress and no public cloud round trip. The system is designed to be useful without any outbound connection, which is what makes it lawful for firms that cannot send data offsite.

What does it mean that the patents are filed rather than granted?

We have 104 filed UK patent applications, roughly 2,340 claims across 13 families, owned by Mickai LTD. Filing establishes priority and a prior-art moat: it fixes the date and stakes the ground. We describe them as filed, never granted, because that is the accurate position.

How can an auditor trust the record without trusting Mickai?

Every consequential action is signed under post-quantum cryptography (FIPS 204 ML-DSA-65 with ML-KEM-768) and hash-chained into an append-only ledger. Anyone can verify that ledger offline, for decades, without trusting the vendor. The proof stands on its own cryptography, not on our assurances.