Sovereign gateway to European data. Patented. We enable physical AI.
The next generation of AI has to learn how things are physically made. That knowledge sits in European factories and will never be sold, because for a manufacturer the process data is the business.
Encryptiq® gives you access to that data and makes it trainable, without the IP ever leaving.
Annual turnover of Europe’s technology industries, from mechanical and plant engineering to electrical and electronics, across 11.7m employees (Orgalim, 2023). Their process knowledge is almost entirely unmonetised.
Mechanical and plant engineering alone: estimated 2025 EU revenue, across 3.1m employees. Decades of process knowledge in every machine, and the first field Encryptiq opens up.
Growth of the industrial AI market, from USD 44bn in 2024 to USD 154bn by 2030 (IoT Analytics). The model is only as good as the knowledge it can reach. That is exactly where Encryptiq comes in.
AI 2.0 is physical Physical AI.
Models that perceive, decide and act in the physical world: robots, machines, production lines and autonomous systems. Trained not on text from the internet, but on how things are actually made.
Public data is exhausted. What the next models lack is the record of how real production actually behaves.
The shift
The frontier is agentic and physical: models that plan, control and act. They need tolerances, parameter windows, defect patterns, maintenance histories and control traces.
The market has started paying. In the US bankruptcy of Spirit Airlines, Google bought one company’s internal records for USD 10m; AI data provider Mercor stood by with USD 7.5m (Reuters, August 2026).
The lock
Those parameters are decades of engineering judgement. Shipping them to a foreign model builder hands over the advantage permanently and irreversibly.
The EU Data Act created a legal right to machine data in September 2025. It did not create the trust needed to pool it. Contracts never solved that. Hardware does.
The inversion
The capability travels. The data does not.
This is the inversion behind Encryptiq. Everything the platform exchanges fits into three objects. Nothing else crosses a boundary, and no raw data appears in any of them.
Job Order
What is computed, on whose data, for whom, and under which licence terms. Agreed before anything is ingested.
Attested Receipt
Enclave measurement, dataset hash, release report, and the results of the memorisation and leakage tests.
Settlement
Payment released against the receipt, split across contributors, and the capability listed in the registry.
The receipt is the oracle. No attestation, no payout. This is also where the patent reads on the live flow: claim 1 describes a sequence executed per request, and the job order is that request.
Three products, one architecture Clearing matches. Enclave computes. Intelligence automates.
Encryptiq® is a registered trademark of Encryptiq GmbH.
Who brings what, who gets what
Supply side
Own the process data
Encryptiq
Clearing Matches supply and demand, sets terms and price. Enclave Computes on the data encrypted, releases only the attested result.Demand side
Need the knowledge, never see the data
Result
A trained model and data insights. The raw data stays where it was.
Released only through the release gate and attested per run. No plaintext ever leaves the enclave.
Patent DE 10 2018 204 447 B4
Claim 1 protects a per-request sequence: local encrypted storage, a key pair generated for each run inside a trusted execution environment (TEE), decryption only there, and the result encrypted to the requester alone.
Granted and in force until 23 March 2038. Protected in Germany, where owners and buyers get this flow only from Encryptiq. Confidential computing hardware itself is open to anyone.
License the patent
Build the protected flow into your own platform or product. Licences are available for the patented architecture, for operators, platforms and technology partners.
License nowRegulatory drivers
Art. 25 requires privacy by design, Art. 32 names encryption explicitly as a security measure. Data that stays encrypted in use, and is decrypted only inside an attested enclave, turns those duties into architecture.
Applicable since 12 Sept 2025. Users of connected products gain access rights to the data their machines generate, which creates legal supply.
Data intermediation services must be structurally neutral. A TEE proves that cryptographically instead of contractually.
Art. 10 data governance and the Art. 53 training-data summary turn proven provenance into a compliance asset.
Mandates trusted execution environments from June 2027.
About Encryptiq A German cybersecurity company built on
ten years of confidential computing experience.
Encryptiq GmbH was founded by Christian G. Junger in 2025. In 2017, Christian and his former teams started developing the core architecture behind what would become a DPMA-granted patent in 2019. Built with exceptional teams across two prior ventures in Confidential Computing and one successful exit later, the technology matured through real-world deployment in some of the most demanding data environments in the world.
Encryptiq holds the patent rights behind the process architecture for secure data processing and Confidential AI execution across hardware-isolated compute environments in multi-stakeholder data ecosystems. The framework enables protected AI models, secure AI agents and encrypted analytics workloads to operate on sensitive data without exposing it.
Our missionBecome Europe’s leading data gateway for physical AI. Every machine that learns from European industry learns through Encryptiq, without the IP ever changing hands.
Awards
ECSO Cyber Investor Days Award 2020
The patented solution, recognised by the European Cyber Security Organisation as one of Europe’s most promising cybersecurity innovations.
OVHcloud Startup Program 2022 · Rising Star Award
For the technology developed through Securitee, an earlier venture building on this patented architecture.
Management
“Confidential AI must be sovereign by design. Sovereignty must be enforceable by architecture. Encryptiq defines the architecture and process governance layer beneath Confidential AI and data ecosystems. Sovereign and made in Europe.”
Memberships
Initiative
Tell us which capability you need.We will tell you whose data you need for it.
Clearing and Enclave are available today.A first scoping call needs no NDA.