The Segment Nobody Started: Sovereign AI for the Firms the Cloud Cannot Reach
Mickai is a sovereign intelligence operating system that regulated businesses own and run inside their own walls. This is the market the public cloud cannot lawfully reach, and the system we built to serve it.
The segment nobody started
Mickai is a sovereign intelligence operating system that regulated businesses own and run inside their own walls. We built it, and it is live today. It runs entirely on the customer's own hardware, on premises and air gapped, with zero data egress and no public cloud round trip. We are describing a working system, not a concept and not a roadmap.
We serve a market that the incumbents structurally cannot address. Around 0.85 million UK businesses, roughly 15 percent of the total, and about 5 million more across the EU, legally cannot send their data to public cloud AI. These are not laggards waiting to be convinced. They are banks, insurers, hospitals, defence suppliers and public bodies bound by rules that make the standard AI delivery model unlawful for their most valuable data. The demand exists. The compliant supply, until now, did not. That gap is the segment nobody started, and we started it.
Why the wall exists
The constraint is not preference or caution. It is written into regulation. PRA model-risk expectations under SS1/23 require firms to evidence how models behave and to reproduce their outputs. UK GDPR governs special category data with duties that a shared cloud tenancy struggles to satisfy. The NHS Data Security and Protection Toolkit sets a floor for patient data handling. The EU AI Act classifies a wide band of financial, clinical and public sector use as high risk, with obligations that attach to the deployment, not just the vendor. ITAR and EAR restrict where defence-relevant data may travel. The NIS Regulations impose security duties on essential services. The US CLOUD Act means data held by a US cloud provider can be compelled by a foreign authority regardless of where the servers sit.
Taken together, these do not merely raise the bar for public cloud AI. For a large class of data, they close the door. We designed for that reality from the first line. Because Mickai runs inside the customer's own walls with no egress, the question of where the data went, and who could compel it, never arises. There is nothing in transit and nothing in a shared tenancy to subpoena.
What we actually built
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 to a regulated buyer more than raw capability. Because the arbiter is deterministic, the same input produces the same output. Reproducibility is not an aspiration we bolt on afterwards. It is a property of the architecture, and it maps directly onto the model-risk evidence a supervisor expects.
Every consequential action is recorded in the Open Audit Record. Each action 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 us as the vendor. This inverts the usual audit relationship. The customer does not take our word that a control fired. They hold cryptographic proof they can check themselves, long after the event, with no dependence on our continued existence.
The system is delivered as studios, each named from Greek and each doing serious work. Nemesis covers fraud and AML. Plutus handles finance and FP&A. Tyche runs underwriting. Prometheus does forecasting. Iris manages customer service. Nomos governs compliance and Astraea covers legal. Panacea addresses clinical work, Pythia business intelligence, and Aletheia audit. Vinis provides voice, the Agentic Marketing Team handles marketing, Trust Agent holds the perimeter, and OAR-as-a-Service exposes the audit record as a product in its own right. Across fielded units, Pantheon, our post-quantum Layer 1 now on testnet, provides multi-node attestation with no central server, so a network of deployments can prove its integrity to itself.
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. Filing is the point that matters here: it establishes priority and builds a prior-art moat around the way we combine sovereign execution, deterministic routing and post-quantum audit. We are a UK company, Mickai LTD, Companies House number 17166618, with Birmingham manufacturing secured. The engineering, the filings and the production base sit together in one place, under one owner.
The size of the opening
The sovereign AI market is roughly USD 40 billion in 2025 and is projected to reach about USD 148 billion by 2032. That growth is not driven by fashion. It is driven by the same regulatory forces that built the wall. As the EU AI Act's high-risk obligations bite, as supervisors sharpen model-risk expectations, and as data sovereignty moves from a procurement footnote to a board-level requirement, the population of buyers who need on-premises, air-gapped, reproducible AI expands rather than contracts. We are positioned at the front of that curve with a system that already meets the standard, not a plan to meet it later.
Two buyers, one stack
Our thesis has two sides. The first is direct: we sell sovereign AI to regulated firms the public cloud cannot lawfully reach. The second is structural: we license the patented stack to the platforms that want to reach those same firms and currently cannot. Internal analysis maps 196 companies and 311 patent-company pairs as potential licensees, including names such as Microsoft, AWS, NVIDIA, Google, Adobe and IBM. We are clear about what that is. It is potential-licensee sizing, not a signed book of business and not an infringement claim.
We frame this as alliance, not attack. We are an ally to the AI majors, not an OpenAI killer. A platform that adds a sovereign layer to its offering instantly reaches a regulated market it cannot serve today, without rebuilding its cloud economics or waiting years for the compliance architecture to mature. We already hold that architecture, filed and running. The strategic logic follows from the facts rather than from any pitch: the fastest route into the regulated segment is the one already built for it.
Where this leaves a serious buyer
For a regulated enterprise, the calculus is straightforward. The workloads currently held back by legal constraint can move onto AI without leaving your walls, without egress, and with an audit record you can verify yourself for decades. Reproducibility, sovereignty and post-quantum integrity are present at the architecture level, which is where supervisors and auditors will look. We built the segment that the market described but did not supply, and we intend to hold the front of it as the regulatory drivers push more of the economy into 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 running everything on our own hardware mean weaker AI?
No. We run about fifty specialist models with cross-model routing under a deterministic arbiter, tuned for the domains regulated firms actually operate in. Determinism gives you reproducible outputs, which is what model-risk evidence requires, and on-premises execution removes the egress and jurisdiction problems that make public cloud AI unusable for your most sensitive data.
What makes the audit record trustworthy if it comes from the vendor?
The Open Audit Record does not ask you to trust us. Each 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. You can verify that ledger offline, for decades, with no dependence on Mickai continuing to exist. The proof sits in your hands, not ours.
Are the patents granted?
They are filed, not granted. We have 104 filed UK patent applications, roughly 2,340 claims, across 13 invention families, owned by Mickai LTD. Filing establishes priority and builds a prior-art moat, which is precisely the protection that matters at this stage.