Negotiators from the Council and Parliament provisionally agreed on May 7, 2026, to amendments to the AI Act as part of the so-called Digital Omnibus. Of particular relevance is that the Machinery Directive is to be exempted from the direct applicability of the AI Act, as it is planned to regulate AI-related health and safety requirements in the future via delegated acts within the Machinery Directive. Brussels aims to avoid double regulation between horizontal AI law and existing product law.

Is this the breakthrough for Europe’s industrial AI?

The EU is shifting the compliance framework closer to product regulation: machines, equipment, robotic systems, and automation solutions are not to be automatically assessed in parallel under the AI Act and sector-specific safety law. Germany’s leading weekly business magazine, “WirtschaftsWoche,” therefore views the agreement as a positive development for the German mechanical engineering sector and notes a palpable sense of relief within the industry. At the same time, it poses the crucial follow-up question: Will this actually be the breakthrough for Europe’s industrial AI—or is it merely the end of a previously convenient excuse?

From Hanover to Brussels

The fact that this issue is politically explosive in the context of industrial policy was already very clear at HANNOVER MESSE 2026. There, industrial AI—particularly “Physical AI” in machines, systems, and robots—was discussed not only as a technological trend but as a matter of competitiveness. According to WirtschaftsWoche, entrepreneurs there appealed en masse to Federal Chancellor Friedrich Merz to stop the flood of new regulations. Merz took up the call and announced his intention to free industrial AI from the “currently too restrictive framework” of EU regulation.

HANNOVER MESSE 2026 had prominently placed industrial AI at the center and identified AI as a key driver for more efficient, resilient, and sustainable production. Physical AI plays a decisive role here—that is, AI systems that interact directly with the physical world, such as in robots, machines, and plants. Thus, the subsequent regulatory relief in Brussels was not merely a regulatory process but the political response to a debate that had already been openly conducted at the world’s most important industrial gathering.

Physical AI as Europe’s Hope for Industrial Policy

The approval comes at a time when physical AI has become Europe’s hope for industrial policy. This refers to applications that do not primarily generate text or images, but rather make factories more productive: quality control, predictive maintenance, robotics, process optimization, autonomous production control, or data-driven service business models. Many companies see Europe’s competitive advantage in deep domain knowledge, an installed machine base, and industrial data that competitors from purely software-based markets cannot easily replicate.

For the industry, the pressure to act is shifting

The agreement also provides some breathing room in terms of timing. For standalone high-risk AI systems, the new implementation date is set for December 2, 2027; for high-risk AI embedded in products, it is August 2, 2028. For manufacturers, this means additional time to robustly establish classification, documentation, risk management, data quality, logging, human oversight, robustness, and cybersecurity. These obligations do not disappear; they are more closely integrated with existing product safety.

For the industry, this shifts the pressure to act. Until now, companies could often cite regulatory uncertainty as a barrier to innovation. Following the agreement, this argument loses its weight. Anyone seeking to scale industrial AI must now classify AI portfolios by risk, product proximity, and safety relevance; make machine data usable; and more closely integrate governance, CE conformity, cybersecurity, and software update processes.

Turning the reduced regulatory burden into a competitive edge

In summary, the new leeway should be interpreted not as a free pass, but as a vote of confidence. Europe’s mechanical engineering sector now has better regulatory conditions to translate AI into real value creation. Whether this results in a competitive advantage is no longer decided solely in Brussels, but in development departments, factories, and data rooms. Companies that now integrate robust industrial AI applications into their products and processes can turn this reduced regulatory burden into a competitive edge.