About Infercom
Infercom is an EU sovereign AI inference provider. We run open-weight models on purpose-built AI chips that we own and operate in a data center in Munich, for companies that build products and processes on AI. We are based in Luxembourg and have been live since December 2025.
Open-weight models behind one API that works with OpenAI and Anthropic clients, paid per token. You switch by changing the base URL and the model name.
Inference API: Inference APICapacity reserved for you on our racks, for teams that need planned throughput and stable response times.
Dedicated capacity: Dedicated capacityThe same platform in your own data center, set up and run together with you.
On-premises: On-premisesWe do inference only. We do not train models.
Infercom started from two ideas. European companies should not have to send their data outside Europe to use strong AI models. And inference, not training, is where most of them spend their AI budget. So we looked for chips built for inference instead of general-purpose GPUs.
We chose SambaNova, because its chips are made for fast inference and because it was looking for a local partner to run them in Europe. We bought the hardware, and we run it ourselves.
We own the racks in Munich. This is not a rented region of a global cloud.
Learn more: Our own hardware, in the EUChoose our EU sovereign models and your data never leaves the EU. Models that run elsewhere are labelled before you use them. Nothing happens behind your back.
Learn more: You decide where your data goesUp to 10x faster than GPU-based inference, measured on our production API.
Learn more: Speed from purpose-built chipsStandard APIs and open-weight models that you could also run elsewhere.
Learn more: Open models, no lock-inThe chips and part of the software stack come from SambaNova, a US vendor. We say so and explain what we do about it.
Learn more: Honest about what we depend onYou talk to the engineers who run the platform, and your needs shape what we build next.
Create an account, get an API key and change the base URL in your code.
cloud.infercom.aiWrite to us through the support portal or by email. We typically respond within 24 hours on business days.
support.infercom.aiMove from pay per token to dedicated capacity or to your own data center, with the same API.
Learn more: Grow without switchingIf you need a finished application rather than an API, one of our partners builds it on Infercom.
Learn more: Work with a partnerThe people who lead and own Infercom.
Founder & Chairman
Tech entrepreneur and investor. Founded Qualist Technology (2009) and serves as Board Member at Tracim Cement. Active angel investor across enterprise software and deep tech.
Managing Director
Deep expertise in AI and cloud infrastructure. Previously at Vodafone, Bertelsmann, and ICT Cloud Computing Services. Leads Infercom's platform development and go-to-market.
Co-Founder
Founder & Managing Partner at Elefund. Early-stage investor with a portfolio spanning AI infrastructure and fintech, including Groq and Robinhood.
Board Member
Background in private equity and M&A. Previously at Merrill Lynch (Private Client Group) and Esas Holding (Strategic Planning & Business Development).
Infercom is owned by two Swiss investment companies and one private investor.
Pakua Capital AGSwiss Single Family Office
Serik KaldykulovSolo InvestorYes, if you choose our EU sovereign models: they run on our own hardware in Munich. A few models in our catalogue run outside the EU, in a region we do not control. They are labelled in the model list and in the API, so you decide before your first request.
Infercom SCS is a Luxembourg company. It is owned by Qualist Investments AG and Pakua Capital AG, two Swiss investment companies, and by the private investor Serik Kaldykulov.
Infercom is a Luxembourg company with no US parent company. Requests to our EU sovereign models are processed on our own hardware in Munich, with no US CLOUD Act exposure.
No. We run inference only and do not train models. Your prompts and outputs are never stored persistently.
SambaNova makes the chips and the software stack. We own the hardware and run the service in Munich, and SambaNova supports parts of the operation.
Create an account in the Infercom Cloud, get an API key and change the base URL in your code. If you prefer to talk first, contact us.
Start with a pilot, scale to production. Record-breaking performance with dedicated enterprise support.