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
- 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.
- 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.
- 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.
- Sovereign AI for Defense Contractors
DFARS, NIST SP 800-171, CMMC: the real compliance vocabulary, and where a desk-scale box fits inside it.
- Sovereign AI for Journalists
Source protection and press-freedom: when self-hosted inference is a tool of journalistic privilege, not a hobby.
- Sovereign AI for SMB Manufacturing
Small-to-mid manufacturing: the IEC 62443 angle and the realistic ROI for a single-box deployment.
- 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.
- 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.
- 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.