For many European technology startups, the road to global growth often leads through the United States.
Founders may create a US holding company, look for American investors and build connections in Silicon Valley. There are practical reasons for doing that, especially when a young company wants access to international capital and investors.
But Berlin-based AI company Langdock has taken a different route.
In September 2026, the fast-growing German startup completed a major change to its corporate structure. It replaced its previous US holding-company structure with a Societas Europaea (SE) registered in Germany, bringing its parent company under European law. The process began earlier in 2026 and reportedly cost the company several million euros.
It is an unusual move for a growing technology company.
And it becomes even more interesting when you look at how quickly Langdock has grown.
The company went from a small Berlin startup founded in 2023 to a business reporting a $50 million annual subscription-revenue run rate in August 2026. Langdock’s own timeline shows that it crossed $1 million in annual recurring revenue in October 2024, $10 million in September 2025, $20 million in March 2026 and $50 million just five months later.
So, what exactly is Langdock building, and why does Germany matter so much to its story?
From Three University Friends to a Fast-Growing Berlin-based AI company, Langdock
Langdock began in Berlin in 2023.
Its founders — Lennard Schmidt, Jonas Beisswenger and Tobias Kemkes — met while studying at CODE University in Berlin. The university itself recently described Langdock as one of the fastest-growing AI startups to emerge from its community.
The original idea came from a problem many companies were beginning to face.
Employees were already discovering tools such as ChatGPT and other large language models. These tools could help with writing, research, coding, summarising documents, and many everyday tasks.
But businesses had a different question:
How can employees use AI without losing control over company information?
That question was particularly important for European companies dealing with strict data protection requirements.
Langdock initially positioned itself as a GDPR-focused alternative for businesses that wanted to give employees access to AI while keeping control over how the technology was used.
Y Combinator, which accepted Langdock into its Summer 2023 batch, describes the company as a platform that helps organisations deploy AI securely, from employee chat to AI-powered workflow automation.
The founders also wanted companies to avoid being locked into one AI model.
Instead of forcing a business to depend entirely on one provider, Langdock built a layer through which companies could use different AI models. That approach was one of the reasons the startup attracted early attention.
The $3 Million Seed Round That Put Langdock on the Map
Langdock’s early growth attracted investors relatively quickly.
In April 2024, the company announced a $3 million seed round led by General Catalyst and La Famiglia, with Y Combinator and other investors also participating.
At the time, the idea was straightforward: companies wanted to experiment with generative AI, but they did not necessarily want to commit themselves to one large language model provider.
Langdock’s platform was designed to sit between businesses and those models.
That gave companies more flexibility while also giving IT teams greater control over access, data and usage.
The company has since expanded considerably beyond its original chat product.
Today, Langdock offers chat, workflows, AI agents, integrations and an API, turning the platform into a broader system for bringing AI into everyday business operations.
What Does Langdock Actually Do?
The easiest way to understand Langdock is to think of it as an AI layer for an entire company.
An employee can use an AI chat interface for everyday work. But companies can also connect Langdock to their existing systems and internal information.
That can include workplace tools, documents, internal knowledge and other business applications.
Companies can then create AI agents for repetitive tasks and build automated workflows.
Langdock says its platform is model-agnostic, meaning businesses can work with multiple leading AI models rather than being tied to a single provider. Its enterprise offering also supports different deployment options, including cloud and on-premises environments.
This is an important part of the company’s business model.
Langdock is not trying to become another consumer chatbot that millions of people casually use at home. Its focus is on helping organisations make AI useful across their workforce.
And that market can be much more complicated.
Large companies need administration, permissions, security controls, compliance, integrations and ways to monitor how AI is being used.
That’s where Langdock is trying to build its position.
Merck Shows How the Platform Can Scale
One of the clearest examples of Langdock’s enterprise use is Merck.
Merck worked with Langdock to develop its “myGPT Suite”, giving employees access to a more advanced AI platform while maintaining enterprise controls.
According to Langdock’s customer case study, the platform grew to more than 33,000 users, with more than 14,000 monthly active users after the new suite was launched. Employees had also created more than 3,000 internal AI agents.
That example gives some context to Langdock’s growth.
The value of an enterprise AI platform isn’t only about having a chatbot. It is also about helping a large organisation introduce AI in a controlled way and allowing employees to create practical applications around it.
For a company with tens of thousands of employees, that can mean everything from everyday writing assistance to specialised internal tools.
Why Langdock Created a US Holding Company in the First Place
The recent move back to Germany doesn’t mean Langdock never wanted a connection with the US.
Quite the opposite.
When the company was growing, having a US-registered parent company helped it attract investors and benefit from the network around Y Combinator.
That structure is not unusual among European startups.
However, Langdock’s actual operations and customer data remained in Germany, according to the company. The US parent did not have employees, infrastructure or access to production systems.
Over time, though, the structure created questions from customers and their legal teams.
Businesses dealing with sensitive information increasingly want to know exactly where their data is handled and what legal jurisdictions may potentially affect it.
US laws, including the Cloud Act, became part of that conversation.
Langdock’s decision was therefore not simply about moving an office.
It changed the legal structure of the company itself.
A European Company Under European Law
Langdock has now reorganised as a Societas Europaea, commonly shortened to SE.
An SE is a European company structure that allows a business to operate under a framework recognised across the European Union.
For Langdock, the change makes its corporate identity more closely match the message it has already been giving customers: European data protection, European operations and a European foundation.
The company told Euronews that the restructuring process started in early 2026 and cost several million euros.
About 80% of Langdock is owned by founders and employees living in the EU, according to the company.
Importantly, Langdock isn’t saying that it wants to become a Germany-only business.
The company still plans to work with customers internationally and raise money from international investors.
So the message is less about turning away from America and more about building a global company while keeping its legal and operational roots in Europe.
Langdock’s Growth Has Been Remarkably Fast
The numbers tell an interesting story.
According to Langdock’s own milestones:
- September 2023: Langdock launches
- October 2024: $1 million ARR
- September 2025: $10 million ARR
- March 2026: $20 million ARR
- August 2026: $50 million ARR
Here, ARR means annual recurring revenue, or more precisely in the latest $50 million figure, an annualised run rate based on subscription revenue.
That distinction matters because it isn’t the same as saying Langdock collected $50 million in revenue during the entire year.
Still, the progression shows how quickly demand for the company’s product has increased.
Its current website says it is used by more than 10,000 companies, while the latest reporting around its corporate restructuring puts the number at about 13,000 organisations.
That is a very different company from the small startup that entered Y Combinator in 2023.
The Bigger Ambition: Building More of the AI Infrastructure
Perhaps the most interesting part of Langdock’s latest move is what comes next.
The company has said it wants to eventually build a sovereign, full-stack AI platform capable of competing with major US hyperscalers over time.
That’s a huge ambition.
Companies such as Amazon Web Services operate at a completely different scale. AWS generated $128.7 billion in revenue in 2025, according to Amazon’s financial results.
Langdock’s $50 million annualised subscription run rate is therefore still tiny compared with the world’s largest cloud platforms.
But Langdock isn’t claiming that it is already operating at that level.
Instead, the company is laying out a long-term direction.
It plans to launch three new services before the end of 2026 and build its own data centre in Germany. The initial facility is expected to support open-source AI models and provide computing capacity, with the possibility of expanding as demand grows.
If that plan develops successfully, Langdock could gradually move from being primarily an enterprise AI software company toward becoming a broader AI infrastructure provider.
Germany’s AI Startup Scene Is Growing Too
Langdock’s story is also happening at a time when Germany’s startup ecosystem is seeing strong AI activity.
According to Germany Trade & Invest, 3,053 new tech companies were founded during the first half of 2026, the highest number recorded for a comparable six-month period. More than one-third — 1,038 startups — had an AI focus.
Germany is also home to a growing group of technology companies working in AI, defence technology, space and other advanced sectors.
The country’s startup ecosystem now includes 38 unicorns, according to the 2026 Global Unicorn Index figures reported by Germany’s federal government website.
That doesn’t mean every European AI company will succeed.
The AI market is highly competitive, and building infrastructure requires enormous amounts of capital, engineering talent and computing power.
But it does show that Europe’s technology story is becoming broader.
What Makes Langdock’s Story Interesting?
Langdock’s decision is interesting because it represents a different idea of what a global technology company can look like.
It doesn’t have to choose between being European and being international.
The company can build in Berlin, operate under European law, work with international investors and sell its products around the world.
Its founders have already shown that the company can grow quickly without abandoning its European base.
The next stage will be more difficult.
Moving from enterprise AI software toward infrastructure is a much bigger challenge. Building data centres, providing computing capacity and competing with established cloud providers require significant resources.
But Langdock has already demonstrated something important: there is demand for AI products that focus on security, flexibility and enterprise adoption.
Its partnership with companies such as Merck gives it a practical environment in which to develop those products, while its growing revenue gives the company more room to invest.
A New Chapter for Berlin-based AI company Langdock
Langdock’s journey began with three university friends in Berlin who saw a simple problem: businesses wanted to use AI, but they needed more control over how it was introduced.
Three years later, the company has thousands of organisational customers, tens of millions of dollars in annualised subscription revenue and a much bigger ambition.
Now it has made another unusual choice.
Instead of keeping its US parent structure as the company grows, Langdock has brought its parent company back under European law in Germany.
That decision does not guarantee success, and the company’s plan to build a broader AI infrastructure business will take years to prove.
But it does make Langdock’s story worth watching.
The company is showing that a European AI startup can think globally without necessarily moving its legal centre of gravity to Silicon Valley.
And perhaps that is the most interesting part of the Langdock story: it isn’t trying to prove that Europe can copy Silicon Valley. It is trying to build something global while keeping its roots firmly in Europe.
For Germany’s growing AI ecosystem, that could become an important story in the years ahead.
