QUICK ANSWER
Cohere (Canada) acquired Aleph Alpha (Germany) on April 24, 2026, creating a combined entity valued at $20 billion — the largest non-US-headquartered frontier AI company. The Series E round is anchored by a $600 million Schwarz Group investment. Both the Canadian and German governments endorsed the deal. The combined company operates under the Cohere brand with dual headquarters in Toronto and Germany, targeting European regulated industries and governments that cannot or will not route sensitive data through US-domiciled AI infrastructure.
Why This Deal Exists
The Cohere-Aleph Alpha merger is a direct response to two converging pressures. First, Europe's AI Act creates compliance requirements that make it difficult for organizations in regulated industries to route sensitive data through API-based AI services controlled by US companies. Healthcare records, financial data, legal documents, and government data all face jurisdictional constraints that the US hyperscalers — OpenAI, Anthropic, Google — cannot easily accommodate without significant structural changes to their infrastructure and data governance models.
Second, the geopolitical environment has sharpened the appetite for technology independence. Canada and Germany signed a Sovereign Technology Alliance earlier in 2026 to deepen collaboration on independent AI capacity. Trade tensions under the Trump administration and a broader reassessment of transatlantic technology dependencies have made government and enterprise buyers in both countries more willing to pay a premium for non-US-controlled AI infrastructure. The Cohere-Aleph Alpha deal is the commercial expression of that demand signal.
Neither Cohere nor Aleph Alpha was individually large enough to credibly fill the sovereign AI role at scale. Cohere had $240 million in ARR and a $7 billion valuation before the deal. Aleph Alpha had €2.7 billion (~$3 billion) in valuation from its November 2023 fundraise. Combined at $20 billion — with Schwarz Group's $600 million anchoring the Series E round — the new entity has the capital base and the political backing to pursue the European government and regulated enterprise market that neither could win alone.
What Each Company Brings to the Merger
| Company |
What It Contributes |
Key Customers / Partners |
| Cohere (Canada) |
Large language model development; $240M ARR; enterprise sales motion; Microsoft strategic partnership; private cloud / on-premises deployment specialization |
Royal Bank of Canada, Fujitsu, LG CNS, Saab (GlobalEye defence MOU) |
| Aleph Alpha (Germany) |
European government relationships; small language model and European language expertise; GDPR compliance architecture; strong brand in European AI sovereignty debate |
German federal government, SAP, Bosch, Schwarz Gruppe, Hubert Burda Media |
Cohere CEO Aidan Gomez described the combination as explicitly complementary: Aleph Alpha's strength in small language models and European languages pairs with Cohere's focus on large language models for enterprise workloads. Aleph Alpha brings 250 employees and deep integration with German public sector clients that took years to build. Cohere brings the commercial revenue base, international sales infrastructure, and the technical depth to build the combined product roadmap.
The deal structure gives Cohere shareholders approximately 90% of the combined entity, with the remaining 10% going to Aleph Alpha shareholders. The combined company operates under the Cohere brand with dual headquarters in Canada and Germany. Schwarz Group's involvement as the Series E anchor investor is significant: Schwarz is one of the world's largest retail conglomerates (owner of Lidl and Kaufland) and operates STACKIT, a cloud and digital infrastructure platform. The combined company is expected to deliver sovereign AI capabilities through STACKIT, making Schwarz Group a deployment infrastructure partner as well as a financial backer.
What Aleph Alpha Had Been Doing Before the Merger
Aleph Alpha was founded in Heidelberg, Germany in 2019 and raised €500 million in a November 2023 round backed by SAP, Bosch, Schwarz Gruppe, and the German federal government. It originally built Luminous, a family of frontier language models designed specifically for European data sovereignty — models that could be deployed on-premises within a customer's own infrastructure, never routing data through US cloud services.
In 2024 and 2025, Aleph Alpha pivoted away from competing at the frontier model level — where it lacked the compute scale to keep up with OpenAI and Anthropic — and toward building AI solutions for the public sector and regulated enterprise. This pivot was commercially sensible but strategically limiting: it repositioned Aleph Alpha as a systems integrator and public sector specialist rather than a frontier AI company, which made the $3 billion valuation hard to sustain as a standalone entity. The Cohere merger resolves this: it gives Aleph Alpha the model development backing it lost by stepping back from frontier development, while giving Cohere the European government distribution it could not build organically.
Aleph Alpha co-founder Jonas Andrulis had departed before the merger, which removed a key voice in the European AI sovereignty debate from the combined entity. The departure was noted by some European AI observers as a symbolic loss for the original mission, though the commercial logic of the merger was broadly seen as sound.
What Sovereign AI Actually Means in Practice
Sovereign AI is not a product category — it is a deployment architecture requirement. An organization has sovereign AI when the AI system runs on infrastructure it controls, processes data within its own jurisdiction, and does not send any data to a third-party cloud operated by a foreign company. For most European government bodies, defense organizations, and regulated financial or healthcare institutions, this is a hard legal or policy requirement, not a preference.
OpenAI, Anthropic, and Google can all offer enterprise contracts with data isolation guarantees, but their infrastructure is fundamentally US-domiciled. A German federal government agency routing sensitive data through Azure OpenAI Service is still running on US-controlled infrastructure, subject to US legal jurisdiction and potential US government access under instruments like the Cloud Act. Cohere's core product design — on-premises deployment in the customer's own environment, or private cloud in the customer's jurisdiction — addresses this at the architecture level rather than through contractual assurances.
What This Means for the US AI Competitive Landscape
The Cohere-Aleph Alpha merger does not threaten OpenAI or Anthropic's consumer or general enterprise businesses. It targets a specific addressable market — European regulated enterprises and governments — where US AI companies face structural constraints regardless of model capability. For that market segment, the combined Cohere-Aleph Alpha entity is now the most credible competitor to US AI at scale.
The deal also signals something broader. Investors are beginning to price the sovereign AI thesis as a distinct and investable category, not just a policy preference. A $20 billion combined valuation backed by Schwarz Group's $600 million and two G7 government endorsements is a credible data point that the addressable market for data-sovereign AI infrastructure is large enough to support independent companies at scale. Watch for similar mergers across the EU — Nordic AI companies, French sovereignty-focused startups, and UK defence AI firms are all potential candidates for the same logic.
Frequently Asked Questions
Is Aleph Alpha still building its own AI models?
Aleph Alpha pivoted away from frontier model development in 2024–2025 to focus on public sector solutions. The combined Cohere-Aleph Alpha entity will use Cohere's model development capability (Command, Embed, and future models) as the model foundation, while Aleph Alpha's European language expertise informs fine-tuning and evaluation for European language markets. Aleph Alpha's Luminous models are expected to be maintained for existing clients but not developed as a competing frontier model line.
Who is Schwarz Group and why does its investment matter?
Schwarz Group is a German retail conglomerate — the parent company of Lidl and Kaufland — with annual revenue over €100 billion, making it one of Europe's largest private companies. Its STACKIT subsidiary operates cloud and digital infrastructure for Schwarz Group's retail operations and is being positioned as a sovereign cloud option for European enterprises. A $600 million investment in Cohere and a plan to deliver sovereign AI through STACKIT makes Schwarz Group both a financial backer and a strategic distribution partner for the combined entity.
How does Cohere's on-premises deployment work?
Cohere's Command and Embed models can be deployed directly on a customer's own servers, private cloud, or a cloud account in the customer's jurisdiction — not via Cohere's own API infrastructure. The customer runs the model; Cohere provides the software licence, model weights, and support. Data never leaves the customer's environment. This is structurally different from OpenAI's API (data processed on OpenAI's infrastructure) or even OpenAI's enterprise tier (data isolated but still on Microsoft Azure, a US-controlled service).
Has the Cohere-Aleph Alpha deal closed?
The deal was announced April 24, 2026 and had not yet closed as of late May 2026. It requires regulatory approvals in both Canada and Germany. The deal has the endorsement of both governments, which should smooth the regulatory path, but formal clearance timelines have not been publicly disclosed. The Series E funding round is also expected to close later in 2026.
Should European enterprises now consider Cohere over OpenAI or Anthropic?
It depends entirely on use case and data type. For internal productivity tools, customer service, and non-sensitive content work, OpenAI and Anthropic offer superior model capability and product surface area. For workloads involving regulated data — healthcare records, financial data, government information, legal documents subject to jurisdiction requirements — the Cohere on-premises deployment model addresses requirements that US-headquartered providers cannot fully satisfy through contractual assurances alone. Most large European organizations will end up with both: US AI for general work, sovereign AI for regulated workflows.