Back to stories
Policy

Merz Pushes EU for Industrial AI Carve-Out Before August Deadline

Michael Ouroumis3 min read
Merz Pushes EU for Industrial AI Carve-Out Before August Deadline

German Chancellor Friedrich Merz opened Hannover Messe on April 19, 2026 with a direct message to Brussels: treat industrial AI differently from consumer AI, or watch European factories fall further behind the United States and China. Speaking at the world's largest industrial fair, Merz said he would push the EU to loosen its artificial intelligence rulebook before the AI Act's main application date on August 2.

A Direct Challenge to the AI Act

"I will push to ease the regulatory burden in the EU on AI and, where possible, to exempt industrial AI from the current regulatory straightjacket that is too tight for AI within the European Union," Merz told attendees. He argued that AI will drive "greater efficiency and productivity, optimised use of resources and, above all, reduced costs" on the factory floor — and that the same rules designed to protect consumers shouldn't apply to machines talking to other machines.

Siemens CEO Roland Busch backed the chancellor from the same stage. "It's complete nonsense to treat industrial and machine data the same way as personal data," Busch said, echoing a complaint German industry has aired repeatedly since the AI Act was finalized.

Why the Timing Matters

The AI Act's risk-tiered obligations kick in in stages, but August 2, 2026 is the milestone when general-purpose model rules and much of the enforcement machinery come online. Merz wants an industrial carve-out locked in before that date, not after, so Europe's manufacturing base isn't subject to the same disclosure, documentation, and high-risk classification obligations that shape consumer-facing AI. For companies serving European customers, this guide to the EU AI Act August 2026 deadline business action plan walks through what to do before the deadline regardless of how the carve-out debate lands.

Germany's Compute Ambitions

The speech came paired with infrastructure commitments. Germany is targeting at least a doubling of domestic data-center capacity by 2030 and roughly a four-fold increase in AI processing capacity — from a baseline near 530 megawatts to roughly 2.1 gigawatts, according to reporting around the announcement. That build-out is meant to give German firms somewhere to run the models they're increasingly asked to deploy.

Industry Case Studies on the Floor

Hannover Messe, which hosts roughly 4,000 exhibitors, doubled as Merz's evidence locker. Microsoft and Krones showcased an AI pipeline that cut fluid-simulation time from four hours to under five minutes. ABB demonstrated a real-time industrial co-pilot ingesting factory data streams. TK Elevator showed AI agents briefing technicians on lift histories before they even arrived on-site. The common thread: these are deployments where the AI is not talking to end users at all, and where the compliance overhead built for consumer chatbots feels mismatched to the risk.

The Political Math

Merz's ask is politically awkward. The AI Act was sold in part as Europe's answer to American and Chinese deregulation, and reopening it weeks before it bites would signal that the bloc is willing to dilute its flagship digital rulebook under industrial pressure. But Germany has struggled to land major frontier-model labs on its soil, and Merz is betting that a narrower industrial rulebook is the price of keeping Siemens, Bosch, and the Mittelstand at the AI frontier rather than watching them relocate workloads to US hyperscalers.

Expect a fight. Consumer groups and members of the European Parliament who shepherded the AI Act through its final votes are unlikely to reopen the text willingly, and any carve-out will have to thread guardrails around worker surveillance, physical-safety handoffs, and autonomous control. But with the August deadline now roughly 100 days away, the pressure on Brussels has a clock attached to it.

Learn AI for Free — FreeAcademy.ai

Take "AI Essentials: Understanding AI in 2026" — a free course with certificate to master the skills behind this story.

More in Policy

Google Signs Classified AI Deal With Pentagon for 'Any Lawful Government Purpose'
Policy

Google Signs Classified AI Deal With Pentagon for 'Any Lawful Government Purpose'

Google has entered a classified agreement allowing the US Department of Defense to deploy its AI models for any lawful government purpose, with non-binding limits on mass surveillance and autonomous weapons.

6 hours ago2 min read
EU Heads to Trilogue April 28 With Plan to Delay High-Risk AI Rules to 2027
Policy

EU Heads to Trilogue April 28 With Plan to Delay High-Risk AI Rules to 2027

Brussels negotiators meet today aiming for a political deal on the AI Omnibus that would push the high-risk AI Act deadline back to December 2027 and lock in firm watermarking rules for synthetic content.

19 hours ago2 min read
State Department Orders Global Push to Warn Allies About Alleged Chinese AI Theft
Policy

State Department Orders Global Push to Warn Allies About Alleged Chinese AI Theft

A U.S. State Department diplomatic cable instructs embassies worldwide to raise concerns with foreign governments about Chinese AI firms — DeepSeek, Moonshot AI, and MiniMax — allegedly extracting and distilling models from American labs.

1 day ago3 min read