Reading path

I'm exploring sovereign AI for my industry

Vertical case-by-case. The honest framing first, then seven industries where on-premises AI hits a different cost-or-compliance calculation than cloud.

9 articles, in reading order

  1. The Engineering Honesty Manifesto

    The lens for everything else: six rules I write the rest of the site under. Read this first if the vertical-specific pieces feel too confident.

  2. What 'Sovereign' Actually Means in 2026 (And What It Doesn't)

    The six-dimension test. Use it before believing anyone's sovereignty claim, including mine.

  3. Sovereign AI for Healthcare: GDPR, HIPAA, and the DGX Spark

    Healthcare under GDPR and HIPAA: which regulatory clauses actually drive on-premises, and which are easier in the cloud.

  4. Sovereign AI for Defense Contractors

    DFARS, NIST SP 800-171, CMMC: the real compliance vocabulary, and where a desk-scale box fits inside it.

  5. Sovereign AI for Journalists

    Source protection and press-freedom: when self-hosted inference is a tool of journalistic privilege, not a hobby.

  6. Sovereign AI for SMB Manufacturing

    Small-to-mid manufacturing: the IEC 62443 angle and the realistic ROI for a single-box deployment.

  7. Sovereign AI for Financial Services

    Financial services under MiFID II, DORA, and GDPR Article 22: where regulation rewards self-hosted inference and where it does not.

  8. Sovereign AI for Law Firms

    Law firms and attorney-client privilege: the architectural property a cloud AI vendor cannot restore, with the work-product and discovery angles.

  9. Sovereign AI for Public-Sector Pilots

    Public sector procurement, IT-SiG 2.0, EU procurement directives: when the audit-trail demand favors open-stack sovereign AI.

← All articles