Most AI deals get framed as scaling plays. This one is different. When Cohere and Aleph Alpha announced their merger in April 2026, both of their governments showed up to the press conference.
That detail tells you everything about what sovereign AI actually means in practice. It’s not a technical concept. It’s a political one.
The deal pairs Canadian AI startup Cohere with Germany-based Aleph Alpha, backed by the Schwarz Group (parent company of grocery chain Lidl). The intent is to offer a sovereign alternative to enterprises in an AI landscape dominated by the Americans.
The Cohere Europe expansion feels like the clearest signal yet that sovereign AI is moving from political talking point to actual infrastructure strategy.
Key Takeaways
- Cohere is merging with Aleph Alpha in a deal valued at approximately $20 billion.
- Schwarz Group is anchoring the deal with $600 million in structured financing.
- Cohere holds roughly 90% of the combined entity, making this closer to an acquisition than a traditional merger.
- The combined company targets highly-regulated sectors: defense, energy, finance, healthcare, and the public sector.
- The deal is a direct response to enterprise demand for AI infrastructure outside US cloud control.
- Both Canada and Germany have formally backed the AI alliance, giving it unusual geopolitical weight.
What Is Cohere and Why Its Expansion Matters
Cohere is an enterprise AI company that builds LLMs for businesses. It focuses entirely on enterprise deployments where compliance, data control, and on-premise options matter.
It’s not the biggest name in consumer AI, but in the world of regulated enterprise, Cohere has been a serious contender. Last valued at $6.8 billion before this deal, Cohere reported $240 million in annual recurring revenue in 2025.
The Cohere Europe expansion matters because it’s the company’s clearest move yet toward becoming the default AI infrastructure layer for organizations. Best for those who won’t or can’t use American cloud providers.
The Cohere and Aleph Alpha Deal Explained
Here’s the structure, because the framing around this deal has been a bit misleading.
Both companies publicly described this as a merger, but the equity split tells a different story. Cohere’s shareholders receive roughly 90% of the combined entity, with Aleph Alpha shareholders receiving the remaining 10%. In substance, this is a Cohere acquisition of Aleph Alpha, politically softened to serve both governments’ interests in framing it as an equitable transatlantic collaboration.
Framing it in this way matters because Germany gets to say it has a sovereign AI champion. Canada gets to say it’s a global AI player. And Cohere gets a foothold in the EU public sector that would’ve taken years to build independently.
Schwarz Group, already one of Aleph Alpha’s main shareholders, committed €500 million (approximately $600 million) in structured financing and is acting as lead investor in Cohere’s Series E round. The valuation has been anchored at around $20 billion.
The combined entity plans to target highly-regulated industries, including defense, energy, finance, healthcare, manufacturing, and telecommunications, as well as the public sector.
What actually makes the Cohere and Aleph Alpha deal interesting from a product perspective is the technical complementarity. Cohere CEO Aidan Gomez noted that Aleph Alpha’s focus on small language models, European languages, and tokenizers complements Cohere’s broader focus on LLMs.
Why Cohere Is Expanding in Europe
Enterprise demand and data sovereignty
European enterprises and governments have a problem that US AI providers structurally can’t solve.
Data sovereignty in the EU isn’t a preference. For large parts of the public sector and regulated industry, it’s a legal requirement under frameworks like GDPR. You can’t process certain categories of government or health data on infrastructure controlled by an American company subject to US law, regardless of where the servers are physically located.
By combining Cohere’s LLM with Aleph Alpha’s track record in Europe’s regulated sectors. The merger will build a dedicated environment for companies that can’t or will not rely on American cloud infrastructure.
This is also why Schwarz Group’s STACKIT platform is part of the deal. The deal leverages the Schwarz Group’s STACKIT cloud platform to ensure the technical stack operates independently of existing US-dominated compute layers. That matters enormously to a German hospital, a French defense contractor, or a government ministry anywhere in the EU. They want an AI infrastructure that they can audit, trust, and that doesn’t route data through American hyperscalers.
The Cohere Europe expansion isn’t just about the market size. It’s about capturing a structural demand that OpenAI, Microsoft, and Google are actively unable to serve for compliance reasons.
How This Signals the Rise of Sovereign AI
Sovereign AI is a concept that’s been floating around policy circles for a few years. The Cohere and Aleph Alpha deal is the first time it’s shown up in a merger announcement with a $20 billion price tag attached.
The merger marks a major shift from model development for regular use to focusing on sovereign AI. This confirms a core signal that technical superiority is no longer the primary differentiator. Control, compliance, and regulatory alignment are now the benchmarks for high-value contracts.
A couple of years ago, the AI conversation was almost entirely about benchmark performance. Who had the best scores on MMLU, who could write better code, etc? Now the enterprise conversation has moved to, ” Who can I actually trust with sensitive data? Who can I run on-premise? Who has a legal entity in my jurisdiction?
AI sovereignty means building AI systems where the data and the compute all stay within a country’s geographic boundaries. The Cohere Europe expansion is a bet that enterprises are ready to pay.
Why Europe Is Leading the Sovereign AI Shift
Regulation and AI independence
Europe has been preparing itself for this moment for years. The EU AI Act, which entered force in 2024, created compliance obligations, and those obligations make GPT-4 a risky option for many firms.
A few numbers worth knowing:
- The EU AI Act covers systems used in critical infrastructure with high-risk classification. It requires documentation, human oversight, and conformity assessments.
- European cloud market share for US firms (AWS, Azure, Google Cloud) is around 60%, which is exactly what the European AI governance is trying to reduce.
This is why AI governance and AI sovereignty have become inseparable in Europe. The regulatory environment created demand. The Cohere and Aleph Alpha deal is the market’s response to it.
Amid growing tensions with the United States, Canada and Germany recently launched a Sovereign Technology Alliance, which directly supports the geopolitical logic behind this deal.
Sovereign AI vs. Traditional AI Models
Here’s a practical comparison for anyone building or buying enterprise AI right now:
| Feature | Traditional AI Models | Sovereign AI Models |
| Data Residency | Typically US-based infrastructure | In-country or regional hosting |
| Regulatory Fit | Requires legal workarounds for EU/GDPR | Built for compliance from the ground up |
| Vendor Control | Centralized; US law applies | Distributed; local jurisdiction |
| Customization | Limited (API access mostly) | Full fine-tuning and on-premise deployment |
| Audit Rights | Minimal | Full transparency and auditability |
| Target Users | Consumer, general enterprise | Government, defense, regulated industries |
| Price Premium | Lower upfront | Higher, justified by compliance value |
The AI sovereignty premium is real, and it’s growing. For organizations in healthcare, finance, defense, or government, the question isn’t affordability but the risk of not having it.
AI Infrastructure Behind Cohere’s Expansion
GPUs and AI compute capacity.
Sovereign AI sounds clean in policy documents. In practice, it’s a hardware problem.
Running AI models locally, at scale, without routing traffic through American hyperscalers requires some serious computing infrastructure. GPUs don’t move fast; the procurement cycles in Europe are slow. And the gap between declaring AI independence and actually having it is measured in data center construction timelines.
The global AI infrastructure buildout is accelerating. AI data center investment globally is projected to exceed $200 billion by 2026.
AI data centers and local hosting
This is where Schwarz Group’s involvement gets genuinely interesting. Most people know Schwarz as the company that owns Lidl. But Schwarz Digits, its IT division, operates STACKIT, one of Europe’s largest sovereign cloud platforms. The STACKIT platform is designed to ensure the technical stack operates independently of existing US-dominated compute layers.
That means Cohere isn’t just acquiring a customer base in Europe. It’s plugging into existing AI infrastructure that already serves European enterprises and has a domestic data residency architecture. That’s years of infrastructure development that Cohere gets without having to build it from scratch.
Why Infrastructure Has Become a Strategic Asset
The broader point here is that AI infrastructure is no longer purely a technical asset. It’s a geopolitical one.
The emergence of a $20 billion cross-border AI entity signals that the “AI as a commodity” era is ending in regulated sectors. Future growth will be defined by AI as infrastructure, where the winning platforms are those that provide auditability, residency control, and political insulation.
Cohere’s bet is that the companies that control sovereign AI infrastructure in Europe will occupy a structurally defensible market position. That’s harder to replicate than a better model score.
The Geopolitics of AI: Europe vs. US vs. China
Europe’s Regulation-First Strategy vs. the US and China
The US approach to AI development has largely been permissive. Minimal federal regulation, heavy private investment, and competitive pressure are pushing capabilities upgradations. China has pursued state-directed development, with sovereign AI infrastructure as a national security priority from the beginning.
Europe has chosen a third path. Regulation-first. The EU AI Act, GDPR, and now emerging AI governance frameworks create a compliance environment. All these slow adoptions create defensible demand for compliant infrastructure.
One relevant data point is that the EU represents roughly 450 million people and one of the world’s largest GDP contributors. If AI regulation makes sovereign infrastructure a requirement for market access in the EU, the market for solutions like Cohere is substantial.
The honest tension here is that Europe’s regulation-first strategy carries a real cost. It slows the adoption compared to the US, and can delay the deployment of AI that might genuinely help people.
Challenges Facing Cohere and Sovereign AI in Europe
Cost, regulation, and competition
I’d be doing you a disservice if I only told the optimistic story here. The Cohere Europe expansion faces real headwinds.
Cost: Sovereign AI infrastructure is expensive. Running models on local, audited, compliant infrastructure costs more than routing API calls to AWS. Enterprises in regulated sectors may want sovereignty in principle, but push back on the cost difference in practice. Cohere needs to close that gap.
Integration risk: The primary failure risk lies in bridging the Canadian and German engineering cultures. Watch for talent attrition in the first 180 days post-merger. Cross-border tech integrations frequently underperform due to alignment failures that show up 6 to 12 months in. Aleph Alpha’s previous CEO, Jonas Andrulis, also departed before this deal closed, which adds organizational uncertainty.
Competition: Mistral AI, the French large language model company, is the most direct competitor for European sovereign AI positioning. It already has deep relationships with European governments and a strong AI governance narrative.
Reports suggest xAI has explored a potential three-way partnership with Mistral and Cursor, though it’s unclear whether Mistral would risk undermining its positioning as an alternative to US tech. If that consolidation does happen in some form, Cohere will face a well-capitalized European competitor with strong local credibility.
Regulatory approval: The deal still needs regulatory sign-off. EU merger review for a $20 billion cross-border AI deal involving government-backed entities is not a formality. It will take time, and conditions may be attached.
What This Means for Businesses and Developers
If you’re building on AI infrastructure or making procurement decisions for an organization in a regulated sector, here’s the practical read:
- Data residency is becoming a must-have, not a nice-to-have. If you’re building in healthcare, finance, or the public sector in Europe, you need a clear answer to where your data lives and under whose legal jurisdiction.
- Sovereign AI pricing will normalize. Right now, fully sovereign deployments carry a cost premium. As more infrastructure comes online and competition increases, that cost will compress.
- Enterprise AI evaluation criteria are shifting. Technical capability still matters, but compliance, auditability, and data control are now first-order requirements in procurement processes for regulated industries.
- Developer ecosystems matter. Cohere’s API and model tooling are mature. Aleph Alpha’s PhariaAI suite adds specialized capability for European language and domain tasks.
- AI governance is a product feature now. If you’re a developer building AI applications for European enterprise clients, the ability to demonstrate compliance with AI regulation is part of your pitch, not an afterthought.
The Future of Cohere and Sovereign AI
The deal closes later in 2026 pending regulatory approval. Assuming it does, the next 12 to 18 months will tell us whether the Cohere and Aleph Alpha merger is the start of a new category or a well-funded bet that doesn’t survive integration.
A few things I’ll be watching:
Watch for EU agencies exclusively mandating only sovereign AI platforms in upcoming procurement RFPs. This step could squeeze out US-based providers from significant government contracts.
Expect a wave of consolidation as mid-sized AI labs seek refuge within established industrial conglomerates to survive the capital-intensive compute race. The Cohere and Aleph Alpha deal may be the first of several similar moves. If it works, it becomes a template.
The broader question for sovereign AI is whether AI governance frameworks in Europe remain stable or are fragmented by member states. France, Germany, and smaller EU countries don’t always align on AI regulation priorities. If the regulatory environment becomes more fragmented, the commercial case for pan-European sovereign AI gets more complicated.
The structural demand is real to be honest. The political backing is unusually strong. But the execution is genuinely hard, and the Cohere Europe expansion is competing in a market where being technically good is no longer enough. You have to be locally trusted, legally compliant, and operationally reliable in environments that don’t tolerate downtime.
Final Thoughts
The Cohere and Aleph Alpha merger isn’t just another AI deal. It’s a $20 billion bet that will make AI sovereignty the dominant frame for enterprise AI procurement in regulated markets.
I think that bet is directionally right. The demand is structural, not cyclical. Governments will continue to push for AI infrastructure that they control. Enterprises in regulated sectors will continue to need sovereign-compliant options. And the US hyperscaler model will continue to create the compliance gaps that players like Cohere are designed to fill.
Whether Cohere specifically wins that market is a different question. The competition from Mistral is serious. The integration risk is real. And the pricing challenge is not yet solved.
But the signal is clear. Sovereign AI in Europe is not a niche trend. It’s the next phase of AI infrastructure, and the Cohere Europe expansion is the most direct evidence we have that the market agrees.
FAQs
Cohere is acquiring Aleph Alpha in a ~$20B merger backed by Schwarz Group’s $600M, supported by the German and Canadian governments.
European enterprises and governments need AI that meets strict data sovereignty and GDPR requirements that US-based providers cannot fulfill.
Cohere takes ~90% of the combined entity. Schwarz Group leads Cohere’s Series E with $600M. The company operates under the Cohere brand with dual HQs.
Mistral AI is the strongest direct competitor, with established EU government relationships and a comparable sovereign AI positioning.
STACKIT is Schwarz Group’s sovereign cloud platform. It gives the combined Cohere entity European-hosted AI infrastructure independent of US cloud providers.
