TL;DR
Thorsten Meyer AI reversed much of its recent case for owning AI infrastructure, arguing that most companies should use the strongest available model with vendor fallbacks. The analysis retains a narrower case for sovereign systems in defense, health care, regulated finance and other legally restricted settings.
Thorsten Meyer AI said in an analysis published July 16 that most companies should prioritize the best-performing AI model over sovereign infrastructure, reversing much of the publication’s recent argument that organizations should own their models and computing stack. The author maintained that sovereign deployment remains justified when law or data classification rules out foreign-controlled services.
The publication said it had spent five weeks making the case for model ownership, local computing capacity and protection from foreign vendors or legal systems. Its new analysis challenged that thesis after eight articles reached similar conclusions, warning that the evidence may have been interpreted through an increasingly fixed editorial position.
The counterargument rests on performance gaps, compliance expenses and delayed product releases. The article cited benchmark results showing an unnamed sovereign model trailing stronger alternatives on software and terminal tasks, but acknowledged that the vendor figures were self-reported and awaiting replication. It also cited prior estimates for staffing, idle computing capacity and certification costs without presenting new independent verification.
The author divided buyers into two groups. Organizations handling classified defense work, national health data or finance covered by DORA-related restrictions may face legal gates that model benchmarks cannot overcome. Most other businesses, the analysis claimed, could obtain adequate resilience by placing a multi-model router in front of external services and maintaining fallbacks.
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?
Capability Gains Versus Control Costs
The argument matters because companies are deciding whether to fund owned AI clusters, certified sovereign clouds and custom model programs or buy access to leading commercial systems. If the capability gap is as large as the analysis claims, choosing weaker models could reduce reliability, slow product development and raise the cost of each successful task.
The opposing risk is that dependence on a foreign supplier can expose customers to service restrictions, price changes and another country’s legal authority. The analysis does not dismiss those risks. It argues that many businesses have priced them too highly while failing to account for the opportunity cost of lengthy qualification and infrastructure programs.
best AI model for enterprise use
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Five Weeks of Sovereignty Arguments
The July 16 article followed a series of Thorsten Meyer AI reports examining European AI providers, cloud ownership, computing resources and foreign influence. Those reports generally favored greater operational control and treated vendor shutdown power as a material business risk.
The new analysis pointed to a reported interruption between June 12 and July 1, when two named models were removed under a Commerce directive and later restored. The author described the 18-day disruption as evidence that vendor restrictions can be survivable when alternatives are available. The source material does not provide the directive, affected customer records or an independent account of the incident.
AI model fallback solutions
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Evidence Behind Cost Claims
It is not yet clear whether the cited benchmark differences would persist under independent testing or reflect performance in each buyer’s workload. The analysis also offers figures including a tenfold certification cost, an 83-times revenue multiple and an €11 billion comparison, but the supplied material does not include the underlying calculations.
The claim that a router can deliver 90% of the resilience for roughly 2% of the cost is also the author’s estimate, not a confirmed industry measure. Switching providers may require prompt changes, security reviews and application testing, while data residency rules can limit which fallback models are available.
cloud-based AI services
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Buyers Face a Legal Test
The analysis urges technology leaders to begin with a legal and data-classification review. Organizations facing binding restrictions would still pursue certified infrastructure, controlled model weights or isolated systems. Other buyers would test multiple model suppliers, measure switching costs and compare the gains from stronger models against the price of greater control. Independent replication of the benchmarks and cost figures will determine how broadly the argument holds.
AI model performance benchmark
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 entirely?
No. The analysis retains sovereignty for legally restricted deployments, including classified work and some health or financial systems. It rejects treating full infrastructure ownership as the default for every organization.
What does using the best model mean here?
It means selecting the model with the strongest measured results for a specific workload, regardless of national origin. The article favors capability and reliability when no legal barrier blocks deployment.
How would a multi-model router help?
A router can send requests to different AI providers and redirect traffic during an outage or restriction. Its effectiveness depends on compatible models, tested fallbacks and permitted data flows.
Are the article’s performance and cost figures confirmed?
Not independently in the supplied material. The benchmark tables are described as vendor-reported, while several cost comparisons come from the publication’s earlier reporting and named outside sources. The figures should be treated as attributed evidence, not settled measurements.
Source: Thorsten Meyer AI
