TL;DR
Germany’s sovereign-AI market is moving from policy statements to funded infrastructure, government procurement and commercial services. The Cohere-Aleph Alpha combination, however, shows that European control remains strongest in data centers and operations, while models and advanced chips still rely partly on foreign suppliers.
German AI company Aleph Alpha announced a combination with Canada’s Cohere on April 24, creating a company reportedly valued at about $20 billion as Germany expands domestic computing infrastructure intended to give businesses and public agencies greater control over sensitive data and AI systems.
The combined company will have bases in Heidelberg and Toronto, according to the source material. Germany’s Schwarz Group is leading a $600 million Series E investment in Cohere, and the companies plan to offer services through Schwarz Group’s StackIT platform. The final division of management authority, intellectual property and model-development decisions has not been detailed.
The transaction comes as Deutsche Telekom and NVIDIA operate an Industrial AI Cloud in Munich containing nearly 10,000 Blackwell GPUs. The system went live on February 4 with about 0.5 exaFLOPS of computing performance. Deutsche Telekom said it raised Germany’s available AI computing capacity by roughly 50%. SAP is the platform partner, while reported early customers include Siemens, Mercedes-Benz, BMW and Perplexity.
Public investment is also increasing. German parliamentary documents cited in the source material allocate €805 million in 2026 toward attracting a European AI gigafactory. SAP, Deutsche Telekom, Siemens, IONOS and Schwarz Group are discussing a joint European Union application. Germany’s SPRIND innovation agency has separately allocated €125 million to its Next Frontier AI laboratory program.
Control Stops Above the Data Center
The spending gives German organizations a way to buy locally operated computing, keep sensitive workloads under German or European law and limit access by outside service providers. Those features matter for defense, intelligence, health care and regulated industries where the location of data and the identity of system operators can shape purchasing decisions.
Yet sovereignty differs across the technology stack. Germany can control facilities, operations and access rules, but the Munich system uses US-designed NVIDIA processors. The Cohere deal also means decisions affecting a leading German model developer may be shared across two countries. The developments support a narrower conclusion: Europe can currently purchase greater operational control, but full technological independence has not been established.
As an affiliate, we earn on qualifying purchases.
Demand Gains a Measurable Price
Consulting firm McKinsey estimated in March that the annual global market for AI services could exceed $1 trillion, with almost $600 billion linked to sovereign AI. That figure is a forecast rather than recorded revenue and depends on how sovereign services are defined. Gartner separately projected European spending on sovereign cloud services at $12.6 billion in 2026, an 83% annual increase.
Government purchasing provides another demand signal. Germany’s domestic intelligence agency reportedly selected France’s ChapsVision instead of US data-analysis company Palantir, while the Bundeswehr excluded Palantir from its cloud projects. These decisions indicate that jurisdiction and supplier control are influencing contracts, although they do not establish the size of the broader market.
“A paradigm shift.”
— Free Software Foundation Europe
Governance and Dependence Stay Unresolved
It is not yet clear how Cohere and Aleph Alpha will allocate board control, research priorities, model ownership or access to training data. The companies’ two headquarters do not by themselves show where decisive authority will sit, which laws will govern particular products or whether European customers will receive independently auditable safeguards.
The scale and schedule of the proposed European AI gigafactory also remain unsettled. Consortium negotiations and funding allocations do not amount to a completed facility. McKinsey’s market estimate, Schwarz Group’s reported plan for as many as 100,000 GPUs and the combined company’s valuation should likewise be treated as estimates or forward plans until contracts, deployments and financial disclosures provide more evidence.
EU Bids and Deal Terms Awaited
Attention will turn to the European gigafactory application, the release of detailed Cohere-Aleph Alpha governance terms and evidence that customers are moving production workloads onto the new infrastructure. Procurement decisions by German and EU agencies will show whether sovereignty requirements are becoming standard contract conditions.
Investors and customers will also watch whether Europe can develop competitive models and semiconductor capacity, rather than only operate imported technology locally. Until those layers change, Germany’s clearest near-term market lies in controlled hosting, secure access and regulatory alignment.
Key Questions
What is sovereign AI?
Sovereign AI generally refers to systems operated under a chosen country’s laws and controls, with defined rules for data location, access and infrastructure. The term does not automatically mean every model or chip was developed domestically.
Did Cohere acquire Aleph Alpha?
The source material describes a combination between the companies, with headquarters in Toronto and Heidelberg and a reported valuation near $20 billion. Full ownership and governance terms have not been disclosed in the supplied information.
Is Germany now independent from US AI technology?
No. Germany has gained more domestic computing and operational control, but major installations still use NVIDIA processors, and several model providers remain based outside Germany.
Who is paying for Germany’s AI expansion?
Funding comes from both sectors. Deutsche Telekom says its Munich cloud was privately financed, Schwarz Group is investing in infrastructure and Cohere, and the German government has allocated €805 million for a prospective gigafactory plus €125 million for AI laboratories.
Source: Thorsten Meyer AI
