MICKAI®ArticlesAstraea for Capital Markets: Sove…
Article · 1 July 2026

Astraea for Capital Markets: Sovereign Legal Intelligence for Deal Teams and Disclosure

Astraea brings legal intelligence to prospectus, disclosure and transaction review inside the firm's own walls, where material non-public information legally cannot touch a public-cloud model.

Author
Micky Irons
Published
1 July 2026
Follow Micky Irons
LinkedInX
Sovereign AIMickaiArtificial IntelligenceOpen Audit RecordPatents
Astraea for Capital Markets: Sovereign Legal Intelligence for Deal Teams and Disclosure

The problem with AI on a live deal

There is a class of legal work that simply cannot be sent to a public-cloud AI service. A draft prospectus. A diligence data room. A working group's mark-up of a sale and purchase agreement. The disclosure schedules behind an equity raise. Every one of these documents is material non-public information, and in most regulated firms the rule is not a preference, it is the law.

Capital markets teams already know this tension. The most useful AI is the AI you can point at the live deal, and the live deal is precisely the thing you are barred from uploading. So deal lawyers and disclosure teams have been left with a choice between speed and compliance. They review prospectuses, verify disclosure and red-line transaction documents by hand, under deadline, because the alternative means routing privileged and price-sensitive material through infrastructure they do not own and cannot audit.

Astraea is built to remove that choice. It is the legal Studio inside Mickai, the sovereign AI operating system, and it runs where the deal already lives: inside the firm, on-premises or air-gapped, with nothing leaving the building.

What Astraea does for a deal team

Astraea is legal intelligence for the work that defines a transaction. It reads the full corpus a deal generates and works across it the way a senior associate would, only faster, and without the document ever leaving the firm's own hardware.

On the disclosure side, Astraea reviews draft prospectuses and offering documents against the firm's own precedent and applicable disclosure standards. It surfaces gaps and inconsistencies between sections, flags forward-looking statements that need risk-factor cover, checks that financial figures quoted in the narrative reconcile to the back-half statements, and traces each material assertion to its supporting source in the data room. Verification, the slow and unforgiving heart of a prospectus, becomes a structured pass rather than a manual scramble.

On the transaction side, Astraea reads the full document set of a deal at once. It compares the latest mark-up against the last clean version, isolates substantive changes from cosmetic ones, maps obligations and conditions across the suite of agreements, and tells a deal team where a negotiated point in the main agreement has not yet flowed through to the ancillaries. It builds and checks disclosure schedules against the representations they qualify. It reads the data room and ties findings back to the diligence questions that matter.

None of this replaces the lawyer. It removes the hours of mechanical cross-referencing that stand between a deal team and its judgement.

Why sovereign is not optional here

For capital markets work, where the AI runs is the whole question.

Material non-public information carries hard legal constraints. Market abuse rules, insider lists, information barriers and the firm's own conduct obligations all assume that price-sensitive material stays inside a controlled perimeter. A general-purpose public-cloud model breaks that assumption the moment a draft is uploaded. The PRA's SS2/21 expectations on outsourcing and third-party risk, UK GDPR treatment of the personal data threaded through every deal, and the cross-border reach of the US CLOUD Act all point the same way for regulated firms: the model has to come to the data, not the other way around.

Mickai is built for exactly that posture. The firm owns and runs it inside its own walls, on-premises or fully air-gapped. Every action Astraea takes is written to the Optimised Audit Record, a tamper-evident, post-quantum-signed log. When a regulator, an auditor, or the firm's own general counsel asks what the AI looked at, what it concluded and when, the answer is a signed record, not a vendor's assurance. That is the difference between an AI you can defend in a supervisory conversation and one you cannot put near a deal at all.

This is the wedge the whole market is waking up to. Roughly 0.85 million UK businesses, around 15 percent, and close to 5 million across the EU are legally constrained from putting their most sensitive work on public-cloud AI. The sovereign AI market is sized at around USD 40 billion in 2025 and on a path to roughly USD 148 billion by 2032. Astraea is one Studio inside the operating system built for that reality.

Part of an operating system, not a point tool

Astraea does not arrive alone. It is one of the Greek-named Studios inside Mickai, alongside Nomos for compliance, Nemesis for fraud and AML, Aletheia for audit, Plutus for finance and Pythia for business intelligence. For a firm running deals, that matters: the legal review, the compliance check and the audit trail share one sovereign substrate and one audit record, rather than three vendors and three perimeters to govern.

The platform underneath is substantial and built. Mickai rests on 104 filed UK patent applications, 2,340 claims, held by Mickai LTD with myself as named inventor. Filed, not granted, which gives a priority position and a prior-art moat rather than a marketing claim. The same IP estate has been mapped against 196 companies and 311 patent-company pairs as potential licensees, including the largest names in the cloud. That is potential-licensee sizing, not signed revenue, and I am careful to frame it as such. But it tells you the category Mickai sits in: sovereign AI infrastructure that a hyperscaler would rather own than compete with. Mickai is an ally to that ecosystem, not a challenger to anyone's model.

In June 2026, a third-party signal underlined that the build is being noticed: Crunchbase ranked me at number four globally, with the Mickai company placed in the top one to two percent. A dated, external data point, nothing more, but a useful one.

The window

Mickai is live and built, and we are building to scale. Birmingham manufacturing is secured, the operating system runs today, and the Studios, Astraea among them, are working software rather than a roadmap.

At this stage we are opening that work to a small number of selected partners: capital markets and corporate teams who feel the public-cloud constraint most acutely and want sovereign legal intelligence inside their own walls before it becomes table stakes. This is a deliberate, narrow window for the right firms, not a broad call.

If your deal team carries material non-public information it cannot put near a public model, that is exactly the problem Astraea was built for. I would welcome the conversation.

Micky Irons, founder and CEO of Mickai. Contact: micky@mickai.co.uk

Frequently asked questions

Can Astraea review a live prospectus without the document leaving the firm?

Yes. Astraea runs inside the firm's own walls, on-premises or air-gapped, as part of the Mickai sovereign AI operating system. Draft prospectuses, disclosure schedules and transaction documents stay on the firm's own hardware and never transit a public-cloud model, with every action written to a tamper-evident, post-quantum-signed audit record.

Why can't deal teams just use a public-cloud AI for this work?

Material non-public information carries hard legal constraints under market abuse rules, insider lists and information barriers, alongside PRA SS2/21 outsourcing expectations, UK GDPR and the cross-border reach of the US CLOUD Act. Uploading a live draft to a general-purpose public-cloud model breaks the controlled perimeter those rules assume. Astraea is built so the model comes to the data instead, keeping price-sensitive material inside the firm.

How does Astraea fit with the rest of Mickai?

Astraea is the legal Studio inside Mickai, alongside Studios such as Nomos for compliance, Nemesis for fraud and AML, Aletheia for audit, Plutus for finance and Pythia for business intelligence. Legal review, compliance and audit share one sovereign substrate and one Optimised Audit Record, rather than separate vendors and perimeters to govern.

Is Mickai a live product or a roadmap?

Mickai is built and live, with Birmingham manufacturing secured, and is building to scale. Astraea and the other Studios are working software running inside the operating system today, not future features.

Subscribe
Get every new Mickai article by email.

Long-form essays on sovereign AI from Micky Irons. One email per article. No tracking, no marketing, no third parties. Every email includes a one-click unsubscribe link.

Prefer RSS? Subscribe at /articles/feed.xml.

Originally published at https://mickai.co.uk/articles/astraea-sovereign-legal-intelligence-for-capital-markets-deals. If you operate in a regulated sector or want sovereign AI on your own hardware, the audit form on mickai.co.uk is the entry point.
More articles
18 Aug 2026
How Telecoms Operators Meet the Telecommunications Security Act With AI That Never Leaves the Network
Telecoms operators meet the Telecommunications Security Act code of practice with AI that runs inside the security-critical boundary on operator-owned hardware. A zero-egress perimeter keeps network configuration and signalling data within operator control, so nothing sensitive crosses out to a public cloud service.
18 Aug 2026
Can energy operators run AI on grid and OT data on-premise to satisfy the Cyber Assessment Framework?
Yes. Energy operators can run forecasting and anomaly detection on grid and OT data entirely on their own hardware, and this satisfies the Cyber Assessment Framework more cleanly than cloud analytics, because telemetry never leaves the audited perimeter and no third-party processor exists to assess.
18 Aug 2026
How Airports Meet EASA Part-IS from February 2026 with On-Site AI
Part-IS applies to aerodrome operators from 22 February 2026 and makes the airport, not its vendor, accountable for information-security risk. Running AI on operator-owned hardware behind a zero-egress perimeter keeps passenger and operational data inside that boundary, so a supplier's SOC 2 cannot discharge it.
18 Aug 2026
Can Automotive Suppliers Use AI on OEM Design Data While Keeping TISAX Prototype Protection?
Automotive suppliers can run AI on OEM design and prototype data and keep TISAX prototype protection, but only when the model runs on their own hardware inside the protected zone. Public cloud AI transmits the data outward, which prototype protection forbids.