TL;DR
Thorsten Meyer AI has challenged its own five-week campaign for AI sovereignty, arguing that most companies may gain more from using the strongest available model with a multi-vendor router. The publication says fully sovereign systems remain justified for organizations facing legal or security barriers, while its cost and performance figures still need independent verification.
Thorsten Meyer AI has challenged its own recent campaign for organizations to own their AI infrastructure, arguing in a July 16 analysis that most companies should use the best-performing model available and rely on multi-vendor routing unless law or security rules require a sovereign system.
The publication said five weeks of reporting had repeatedly favored model ownership, local infrastructure and protection from foreign control. Its new analysis re-examined that position and concluded that full AI sovereignty can impose a large cost and capability penalty on organizations that do not face a legal barrier to using foreign providers.
Its case rests partly on reported benchmark gaps. The article cited 77.6% versus 95.0% on SWE-bench and 63.8% versus 89.5% on Terminal-Bench for models it called Inkling and Fable 5. The publication acknowledged that these figures came from vendor tables and Artificial Analysis, were partly self-reported and were awaiting independent replication.
The analysis also cited higher certification, staffing and infrastructure costs for sovereign deployments. It claimed that SecNumCloud qualification can cost far more than ISO 27001, specialist staff may cost $75,000 to $100,000 a year, and idle self-hosted capacity can carry a steep penalty. These numbers were drawn from the publication’s earlier reporting and named industry sources, but no underlying calculations were included in the supplied article.
Against sovereignty: the strongest case for just using the best model
This publication has spent five weeks arguing one thing — and every piece converged. That should bother you. It bothers me. When eight analyses reach the same verdict, you’re not running an analysis. You’re running a thesis, and the evidence has started arriving pre-sorted.
So here’s the case against — argued properly, with the same evidence, turned around. Not a strawman erected to be knocked down. The version a smart CTO would put to me across a table, and which I have not yet answered in public. The claim: for almost everyone, sovereignty is an expensive hedge against a risk they’ve mispriced — and the rational move is to use the best model and get on with it.
Defence · classified · national health data · DORA-bound finance. The foreign-legal-order risk isn’t theoretical and isn’t insurable by other means — it’s a legal gate. No benchmark opens it. Your alternative isn’t a worse model; it’s no deployment at all.
Statistically, you are. You have a reasonable, politically legible, entirely unbudgeted feeling — and an industry built to monetize it. The capability compounds, the tax is real, the opportunity cost is brutal, and 18 days is survivable.
I’ve spent five weeks arguing you should own your stack. The strongest case against says: for most of you, that’s an expensive way to be worse, sold by people whose real product is a feeling. And that case is mostly right. What survives is smaller and sharper — everything above the router line (the qualification programme, the owned cluster, the custom pre-training run, the €11B data centre) you should buy only if a law requires it, never because a narrative does. A router is the sovereignty most people actually need. 90% of the resilience for ~2% of the cost — and it would have made 12 June a non-event. So run the honest test: are you bound, or are you performing?
Model Quality Versus Sovereign Control
The argument matters because technology leaders must decide whether control over models, data location and infrastructure is worth accepting higher costs, slower deployment or weaker model performance. For companies competing through software, delays during qualification and infrastructure work may also mean lost product releases and customers.
Thorsten Meyer AI now draws a line between organizations that are legally bound and those making a discretionary risk choice. It places defense, classified workloads, national health data and some DORA-covered financial services on the sovereign side. For those users, foreign jurisdiction or deployment restrictions can block adoption regardless of benchmark performance.
For other organizations, the publication recommends a router connecting multiple providers, allowing workloads to move when one model becomes unavailable. That approach may reduce dependence at a fraction of the cost of an owned cluster, although the article’s claim of 90% of the resilience for about 2% of the cost is an estimate, not a verified industry standard.
HP ProDesk 4 G1i Mini Desktop PC (Latest Model) Intel Core Ultra 5 235T AI Ready 512GB SSD 16GB RAM WiFi 6E Windows 11 Pro
Overview | Power big ideas in small spaces with a single cable and AI The HP ProDesk 4…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Five Weeks of Pro-Sovereignty Reporting
The July 16 article followed eight analyses in which the publication had argued that customers should examine model ownership, corporate shareholders, computing capacity and the possibility that a supplier could withdraw access. The author said the repeated conclusions had begun to resemble a fixed thesis rather than open-ended reporting.
The reassessment pointed to an alleged service restriction between June 12 and July 1. According to the publication, a Commerce directive removed access to Fable 5 and Mythos 5 for 18 days before both returned, while alternative models remained available. The supplied material does not identify the agency, affected customers or documentary record behind that account, so the episode cannot be treated here as independently confirmed.
“Use the best model. Router in front. Spend the difference on shipping.”
— Thorsten Meyer AI
multi-vendor AI routing software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Evidence Behind the Cost Claims
Several central claims remain unresolved. The benchmark results are not yet independently replicated, and the supplied source does not provide enough methodological detail to compare model versions, test conditions or production reliability. The cost ratios may also vary sharply by workload, utilization, security level and contract terms.
It is also unclear how many organizations truly face binding sovereignty requirements. The publication contrasted a CISPE figure saying 72% cite sovereignty with a Gartner view that it determines purchases in only three verticals. Those measures may cover different populations or definitions, and the supplied material does not include the original studies needed to reconcile them.
DULIWO Scribing Tool Kit for Gunpla – 7-Blade Model Scriber Chisel Set (0.1–2.0mm), Pin Vise Hand Drill with 10 Bits, Tweezers & Brush Included for Gundam HG/RG/MG, Resin Kits, Panel Line Engraving
Model Kit Tools: Includes metal scribe tool (1 handle + 7 blades), small hand drill set (10 bits…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Procurement Tests Move to Risk
Organizations adopting the article’s framework would next need to document whether a law, regulator or security classification prohibits foreign model use. Those without such a barrier could test provider routing, fallback quality, data-handling controls and outage recovery before committing to dedicated infrastructure.
The argument will depend on whether independent benchmark testing supports the reported capability gaps and whether detailed cost studies confirm the claimed sovereignty premium. Buyers will also need evidence that routers can meet their privacy, latency and continuity requirements under real operating conditions.
AWS CERTIFIED Generative AI Developer Professional Certification Exam: A Practical Study Guide
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
Is Thorsten Meyer AI abandoning AI sovereignty?
No. The publication now supports sovereign deployments for legally restricted workloads, including classified, defense, national health and some regulated financial uses. It questions broad adoption by companies without a binding rule.
What does using the best model mean?
It means selecting a model based mainly on capability and business performance, even when the provider or infrastructure is foreign. The phrase does not identify one permanent winner, since model rankings and workload needs can change.
How would a model router reduce dependence?
A router can direct requests among multiple AI providers and move traffic when one service is unavailable. Its protection depends on compatible applications, tested fallbacks and providers that do not share the same legal or infrastructure risk.
Are the benchmark and cost numbers confirmed?
They are reported figures, not fully verified findings. The publication says some benchmark data are self-reported and awaiting replication, while the supplied article does not show the full calculations behind its cost comparisons.
Which organizations may still need sovereign AI?
Organizations handling classified information, defense systems, protected national health data or workloads barred from foreign legal control may have no compliant alternative. Each buyer must establish the exact rule with its legal, security and regulatory teams.
Source: Thorsten Meyer AI
