Finding a fault in 14 vehicles instead of 1.4 million
In the traditional setup, quality data moves between an OEM and its suppliers manually, through isolated portals and spreadsheets. Suppliers have no access to fleet data, so they cannot see how their own components behave in the field. When something goes wrong, parts get shipped between locations for investigation, a practice expensive enough to have earned its own name in the industry: parts tourism.
The BMW Group, DENSO, Bosch and several other suppliers have been running the quality management use case standardized within Catena-X, in which the OEM shares field data directly with suppliers who combine it with their own production data. The reported results are unusually concrete. Errors are detected on average four months earlier than through conventional methods. One suspected fault in a camera system was narrowed from a potential 1.4 million vehicles down to 14. Connecting a new supplier to the point of productive data exchange takes about five weeks rather than several months.
The four months matter more than they first appear. Detection speed is cost. Every month a fault propagates undetected adds vehicles to the affected population, and every one of them is a potential warranty claim.
Primary data instead of industry averages
A second case points the same way from a different angle. Product carbon footprint calculations in automotive have long depended on industry averages, because primary data from deeper supplier tiers was either unavailable or locked in incompatible tools.
Ford, Flex and Micron exchange footprint data across three tiers using a shared exchange standard and calculation rulebook. Micron, as the tier-2 supplier, provides primary data to Flex, which incorporates it and passes enriched values to Ford. What makes this notable is not the volume of data but the structure: all three companies use different certified software, chosen to fit their own IT landscapes, and the exchange works anyway.
What the cases have in common
Read together, the two examples share three structural features, and none of them is a technology.
The first is agreed semantics. Data arrives meaning the same thing to the receiver as it did to the sender, because the models and calculation rules were negotiated in advance rather than reconciled afterwards.
The second is retained control. Each participant decides what to share, with whom and under what conditions. This is what makes it possible for direct competitors and for partners of very unequal size to exchange operational data at all.
The third is freedom of implementation. Each participant connects with software of its own choosing and still exchanges data with the rest. This is where a data space differs from a customer portal: nobody has to adopt another company's system, and nobody is locked into a single provider.
Why value alone does not produce adoption
Every one of these benefits scales with participation. The costs do not. A supplier that connects pays the full onboarding effort itself, while what it gets back depends on how many of its customers and sub-suppliers are also connected. Field data reveals a pattern only when enough of the fleet is covered. Footprint values become more accurate only when the deeper tiers are actually there to supply them.
Moving first therefore means carrying cost for a benefit that arrives only once others follow. A company can be entirely convinced by the numbers above and still rationally wait. This is a coordination problem, and evidence alone does not resolve it: the bottleneck is neither persuasion nor architecture, but the gap between individually borne cost and collectively realized value in the period before critical mass.
Measuring what happens when the gap is closed
This is the reasoning behind the Data Space Accelerator, a study coordinated by IDSA together with the Catena-X Automotive Network e.V. and Cofinity-X, funded by the German Federal Ministry for Economic Affairs and Energy. Its design is worth noting because it is a study rather than a subsidy program. Companies receive structured support to reach productive data exchange, and in return they take part in accompanying research on what that exchange is actually worth to them.
The research question is the one left open by the cases above. Individual implementations demonstrate value. What has never been measured at scale is what happens to that value as participation grows across an entire value chain, including the smaller suppliers who are usually absent from lighthouse projects and who carry the highest relative onboarding burden.
If the answer is what the current evidence suggests, the argument for industrial data sharing stops depending on models. Details of the study, including eligibility and the accompanying documentation, are published at https://data-space-accelerator.com
At Data Expo
Dr. Mario Holesch will present Data Sharing: Economic Value Across the Automotive Supply Chain, Built on Trust at Data Expo, Lezingenzaal 2, on Wednesday, September 9, 2026, from 16:30 to 17:00 at Jaarbeurs Utrecht.
This blog post is a contribution from the International Data Spaces Association (IDSA), an international organization that invented the concept of data spaces and develops the global standard for them. For more information, visit www.internationaldataspaces.org or stop by the International Data Spaces Association booth at Data Expo.
