Publications

Almost a year ago, I published a post in German on the EU Data Act — specifically on how manufacturers of connected products might avoid the obligation to share product data.

For someone like me, who has spent years working on the legal frameworks for sharing machine-generated sensor data in data spaces, this was an unusual angle to take.

But here's the thing: trade secret protection is one of the central legal challenges in data sharing. Traditional legal safeguards — like confidentiality agreements — only rarely lead to enforceable damage claims, and even then, compensation is at best a consolation prize when valuable operational know-how has already been compromised.

Given these limited legal remedies, we had previously identified technical approaches like "Compute to Data" as a promising solution. The idea: data stays with the data provider, and only the computational task is sent there and executed locally — keeping sensitive data out of third-party hands.

Then came the EU Data Act, which mandates data disclosure — effectively ruling out Compute to Data — while offering only confidentiality agreements for trade secret protection; i.e., the confidentiality mechanism identified as being inadequate.

This prompted me to explore whether manufacturers could find a legitimate way to avoid the disclosure obligation altogether under the EU Data Act, as a means of protecting trade secrets. This should not be read as a call to withhold data in the context of Manufacturing-X — the opportunities that data sharing unlocks are far too significant for that.

The topic remains as relevant as ever, even though the EU Commission's so-called "Digital Omnibus" proposes some tightening of trade secret protections under the Data Act in favour of manufacturers. But whether and how this will be implemented — and what requirements manufacturers will need to meet to lawfully refuse data disclosure — remains to be seen.

(My apologies to my German-speaking network for the partial overlap with last year's post.)