AI News, Week 29: Most German AI Providers Use U.S. Models
Also: Products with AI components generate over 120 billion euros in revenue, and Claude Cowork keeps working even when the laptop is closed
The Week in 30 Seconds
The AI News for Week 29 shows where companies in Germany stand when it comes to AI: More than half of German AI providers rely on models from the U.S., products incorporating AI generate over 120 billion euros in revenue, and Claude Cowork now continues to operate in the background.
Two of the news stories examine the same question from different angles: How dependent are companies on individual model providers, and where does value come from? Here, you can read about what’s behind the three most important AI news stories of calendar week 29 and what they mean for your company.
GreenPT Analysis: 53 percent of German AI providers rely on U.S. models, while only 10 percent use Mistral
European hosting won’t solve the problem of dependence as long as model access and pricing are determined in the U.S. Companies are only truly independent when they can switch models and providers as needed.
Robin Reuschel
Expert in AI Strategy & Organization
European hosting won’t solve the problem of dependence as long as model access and pricing are determined in the U.S. Companies are only truly independent when they can switch models and providers as needed.
Robin Reuschel
Expert in AI Strategy & Organization
What This Means for Businesses
- Evaluate dependencies on a case-by-case basis: Not every application requires a European model. Determine which dependencies are acceptable for each process.
- Check for interchangeability: Determine whether models are interchangeable, how data exports work, and what alternatives are available in the event of price changes or outages.
- Question provider claims: For software advertised as “hosted in Germany,” ask which model is actually used and who operates it.
To learn what you should consider when selecting AI tools from a compliance perspective, read our article on AI tools and compliance.
Study: Products Incorporating AI Generate Over 120 Billion Euros in Revenue for Companies in Germany
Your competitors also have access to models. Your competitive edge comes from your own data and process knowledge, which you reliably integrate with AI.
Robin Reuschel
Expert in AI Strategy & Organization
Your competitors also have access to models. Your competitive edge comes from your own data and process knowledge, which you reliably integrate with AI.
Robin Reuschel
Expert in AI Strategy & Organization
What This Means for Businesses
- Adapt existing models: Developing a proprietary base model is worthwhile for very few companies. Invest in adapting existing models to your processes and quality requirements.
- Make data usable: Operational and product data are your competitive advantage. Clarify access, quality, and responsibilities before implementing use cases.
- Plan for multiple models: An architecture with interchangeable models reduces your dependence on any single model and allows you to choose the right model for each task.
Our page on AI in the automotive and mechanical engineering industries shows how industrial companies are using AI in production and engineering.
Claude Cowork is coming to smartphones and keeps working even when the laptop is closed
The main change is that AI-driven work is no longer handled in an open session. Before you hand over actual workflows to assistants, you therefore need clear permissions and approvals.
Robin Reuschel
Expert in AI Strategy & Organization
The main change is that AI-driven work is no longer handled in an open session. Before you hand over actual workflows to assistants, you therefore need clear permissions and approvals.
Robin Reuschel
Expert in AI Strategy & Organization
What This Means for Businesses
- Identify tasks related to work: Cost reconciliations, contract overviews, or briefing documents involve many small steps across multiple systems. This is where assistants like Cowork Teams provide immediate relief.
- Limit permissions: Connect only the applications needed for a task. Specify which actions require approval.
- Manage mobile approvals: Anyone approving results on the go needs the same rules as in the office. Clarify which approvals are permitted via smartphone.
Our article on Agentic AI describes how companies are introducing AI agents in a controlled manner.
How independent is your AI landscape?
We’ll work with you to evaluate models, providers, and migration options, and develop a robust AI architecture.
Sources
- GreenPT: German AI Companies and Their Model Providers, July 4, 2026
- t3n: Europe’s AI Problem: Only One in Ten German Companies Uses a European Model, July 13, 2026
- IW Cologne: The Germany Case: How AI Is Transforming the Industrial Business Model, July 2026
- BigData-Insider: Study: AI Innovations Generate 120 Billion Euros in Revenue, July 13, 2026
- The Next Web: Anthropic’s Claude Cowork moves from the laptop to your phone, July 7, 2026
- Claude Help Center: Use Claude Cowork on the web, desktop, and mobile, as of October 2026
Researched with AI assistance, editorially reviewed by Robin Reuschel.
Robin Reuschel
Managing Director ROOVER
Robin Reuschel ist Experte für KI-Strategien bei Roover mit langjähriger Beratungserfahrung bei u.a. Roland Berger. Mit tiefer Expertise unterstützt er Unternehmen bei der KI-Transformation für zukünftige Geschäftsmodelle
You might also be interested in
Chatbots are now everywhere: on websites, in customer portals, in apps. Almost all of us have seen a small chat window asking: “How can I help you?” Nevertheless, many companies still have an uneasy feeling. Is this really good customer service or just another digital filter in front of the
AI has arrived in marketing, but unfortunately often as a wild zoo of individual tools, test accounts and “prompt snippets” in Slack. The result is rarely strategic: more output, but no clean quality, no repeatability, no protection against data and reputational risks. If you want to use AI tools in
AI governance in the company: The practical guide AI is already in use in many teams, but often without uniform rules. This is exactly where AI governance comes into play: who is allowed to do what, what data is permitted, how do we document decisions? Without this, it gets expensive: