Sizing the Sovereign AI Market: From USD 40B in 2025 to USD 148B by 2032

The number that matters is not the model, it is the firm that cannot use it

Most AI market sizing starts with capability and works backward to revenue. I think that is the wrong end of the telescope. The decisive variable in sovereign AI is not how good the model is. It is how many organisations are legally barred from sending their data to someone else's cloud to use one.

That population is large, it is already defined by law, and it is growing. Sovereign AI is forecast to move from around USD 40B in 2025 to around USD 148B by 2032. That curve is not a bet on hype. It is the financial shadow cast by regulation that already exists.

I am Micky Irons, founder and CEO of Mickai. Mickai is a sovereign AI operating system, or SIOS: AI that regulated businesses own and run inside their own walls, on-prem and air-gapped, with every action written to a tamper-evident, post-quantum-signed audit record we call the OAR. It is built and live, not a concept. This piece is about the demand side, the reason we built it the way we did.

Who legally cannot use public-cloud AI

Start with the wedge. We estimate roughly 0.85 million UK businesses, about 15 percent of the active base, and roughly 5 million EU firms sit in a category where routing regulated data through a public-cloud AI service is not a procurement preference. It is a compliance failure.

This is not one rule. It is a stack of them, and most regulated firms are caught by several at once:

• **PRA SS2/21** sets outsourcing and third-party risk expectations for UK financial firms, including data location, exit, and genuine control over critical services.
• **UK GDPR special-category data** raises the bar sharply for health, biometric, and other sensitive processing, where lawful basis and data residency are not optional.
• **NHS Data Security and Protection Toolkit** governs how health and care organisations and their suppliers handle patient data.
• **EU AI Act high-risk obligations** impose documentation, logging, and human-oversight duties on the very use cases that regulated firms most want to automate.
• **ITAR and EAR** treat certain technical data as export-controlled, which makes uncontrolled cloud processing a legal exposure rather than an IT choice.
• **NIS Regulations** set security and incident duties for operators of essential and digital services.
• **The US CLOUD Act** is the quiet decider. It means data held by a US-linked provider can be reachable by US legal process regardless of where the server sits, which is precisely the assurance a European regulator will not accept.

Read together, these do not say AI is forbidden. They say the data cannot leave. That single constraint is the whole market.

Why the constraint creates a market rather than a workaround

The instinct is to assume firms will engineer around this with redaction, tokenisation, or a regional cloud region. Those help at the margin, but they do not resolve the core problem, which is custody. If a third party can technically access the data, or can be compelled to, the firm has not actually retained control, and the auditor knows it.

Sovereign AI is the structural answer. The model runs where the data already lives, inside the firm's own boundary, with no outbound dependency at inference time. That is why the category grows as a function of regulation rather than as a function of model quality. Every new compliance obligation widens the set of workloads that cannot be outsourced, and therefore widens the addressable base for AI that the customer owns and runs.

This is also why I frame Mickai as an ally, not an OpenAI killer. There is a dual-buyer reality here. Plenty of unregulated work is well served by public-cloud AI and will stay there. The sovereign segment is the part that legally cannot move, and that is the part we serve. Two buyers, two deployment models, one growing pie.

From a constraint to a product surface

A market this defined still needs something to buy. Mickai turns the constraint into a working operating system rather than a compliance promise. The capability lives in Greek-named Studio modules, each a domain the regulated buyer already has a budget line for: Nemesis for fraud and AML, Plutus for finance and FP&A, Tyche for underwriting, Prometheus for forecasting, Iris for customer service, Nomos for compliance, Astraea for legal, Panacea for clinical, Pythia for business intelligence, and Aletheia for audit. Around them sit Trust Agent, the AMT marketing system, Vinis voice, and OAR-as-a-Service.

The common thread is the OAR. Every action a Studio takes is written to a tamper-evident, post-quantum-signed audit record. For a regulated buyer that is the difference between an AI they can defend to a supervisor and one they cannot put into production at all. The forecast to USD 148B assumes buyers who need exactly this, and they are the buyers the regulation manufactures.

The IP position behind the thesis

Owning the substrate matters when the substrate is the moat. Mickai holds 104 filed UK patent applications, around 2,340 claims, owned by Mickai LTD with me as inventor. These are filed, not granted, and I am precise about that. Their job at this stage is to establish priority and a prior-art moat around sovereign, auditable, on-prem AI.

That filing base also maps onto a licensing surface. Our analysis identifies 196 companies and 311 patent-company pairs as potential licensees, including names like Microsoft, AWS, NVIDIA, Google, Adobe, and IBM. To be clear about what that is and is not: it is potential-licensee sizing, a read on where the architecture is relevant, not an assertion that anyone is infringing. The dual-buyer thesis runs here too. The same hyperscalers we treat as allies on the deployment side are the firms most likely to need sovereign building blocks as their regulated customers demand them.

A momentum signal, dated

Markets move on conviction as well as on regulation, so one external read. As of June 2026, I am ranked number 4 on Crunchbase by CB Rank Person, verified live, with the Mickai company profile in the global top 1 to 2 percent. I treat that as a third-party momentum signal from a specific moment, not a permanent claim. It tells me the thesis is landing while the underlying regulatory demand keeps compounding underneath it.

We are a UK company with Birmingham manufacturing secured, and we are building to scale.

The opportunity, plainly

The forecast from USD 40B in 2025 to USD 148B by 2032 is, in the end, a count of organisations the law will not let move their data, multiplied by their appetite to use AI anyway. Mickai exists to serve exactly that count, and the platform is built and live today.

A pre-seed window is open to a small set of selected partners as we scale. This is an invitation to get involved early in a sovereign AI category with a regulatory tailwind, not a sign of need. If the thesis fits how you see the next decade, I would value the conversation.

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

FAQ

What is a sovereign AI operating system (SIOS)? It is AI that a regulated business owns and runs inside its own walls, on-prem and air-gapped, with no outbound dependency at inference time. Every action is written to a tamper-evident, post-quantum-signed audit record, the OAR. Mickai is a SIOS, and it is built and live today.

Why can roughly 0.85 million UK firms and 5 million EU firms not use public-cloud AI? Because a stack of existing rules requires that their data does not leave their control. PRA SS2/21, UK GDPR special-category obligations, the NHS Data Security and Protection Toolkit, EU AI Act high-risk duties, ITAR and EAR, the NIS Regulations, and the US CLOUD Act each push regulated data away from public-cloud AI. For these firms routing regulated data offsite is a compliance failure, not a preference.

How large is the sovereign AI market? Sovereign AI is forecast to grow from around USD 40B in 2025 to around USD 148B by 2032. That growth tracks regulation rather than model quality, because every new compliance obligation widens the set of workloads that legally cannot be outsourced.

Is Mickai competing with the major AI labs? No. Mickai is an ally, not an OpenAI killer. The dual-buyer thesis recognises two markets: unregulated work well served by public-cloud AI, and the regulated segment that legally cannot move, which is the part Mickai serves. The hyperscalers are also potential licensees of sovereign building blocks.

What is the state of Mickai's patents? Mickai LTD holds 104 filed UK patent applications, around 2,340 claims, with Micky Irons as inventor. They are filed, not granted. Their role is to establish priority and a prior-art moat around sovereign, auditable, on-prem AI.

How can an investor get involved? A pre-seed window is open to a small set of selected partners as Mickai scales. It is an invitation to get involved early in a category with a regulatory tailwind. Reach Micky Irons directly at micky@mickai.co.uk.

Frequently asked questions

What is a sovereign AI operating system (SIOS)?

It is AI that a regulated business owns and runs inside its own walls, on-prem and air-gapped, with no outbound dependency at inference time. Every action is written to a tamper-evident, post-quantum-signed audit record, the OAR. Mickai is a SIOS, and it is built and live today.

Why can roughly 0.85 million UK firms and 5 million EU firms not use public-cloud AI?

Because a stack of existing rules requires that their data does not leave their control: PRA SS2/21, UK GDPR special-category obligations, the NHS Data Security and Protection Toolkit, EU AI Act high-risk duties, ITAR and EAR, the NIS Regulations, and the US CLOUD Act. For these firms routing regulated data offsite is a compliance failure, not a preference.

How large is the sovereign AI market?

Sovereign AI is forecast to grow from around USD 40B in 2025 to around USD 148B by 2032. That growth tracks regulation rather than model quality, because every new compliance obligation widens the set of workloads that legally cannot be outsourced.

Is Mickai competing with the major AI labs?

No. Mickai is an ally, not an OpenAI killer. The dual-buyer thesis recognises two markets: unregulated work well served by public-cloud AI, and the regulated segment that legally cannot move, which is the part Mickai serves. The hyperscalers are also potential licensees of sovereign building blocks.

What is the state of Mickai's patents?

Mickai LTD holds 104 filed UK patent applications, around 2,340 claims, with Micky Irons as inventor. They are filed, not granted. Their role is to establish priority and a prior-art moat around sovereign, auditable, on-prem AI.

How can an investor get involved?

A pre-seed window is open to a small set of selected partners as Mickai scales. It is an invitation to get involved early in a category with a regulatory tailwind. Reach Micky Irons directly at micky@mickai.co.uk.