BRUSSELS BACKS DOWN: EU GIVES AI COMPANIES 16 MORE MONTHS TO COMPLY WITH STRICT RULES The European Union has pulled back from one of the most consequential enforcement deadlines in the AI Act, agreeing to extend the compliance window for high-risk AI systems by up to 16 months. The Council and European Parliament reached a provisional agreement on May 7 to delay enforcement of the high-risk rules until the Commission confirms that the necessary technical standards and auditing frameworks are in place. The existing deadline of August 2, 2026 would have required companies to comply with sweeping obligations around transparency, human oversight, and risk management before those standards had been fully written, let alone adopted. Industry groups had spent months warning that the infrastructure for compliance simply did not exist. The Commission finally listened. The extension applies specifically to the high-risk AI tier, which covers systems used in hiring, credit scoring, law enforcement, critical infrastructure, and healthcare. This is the portion of the AI Act that global companies feared most and lobbied hardest against. The rest of the AI Act remains on its original timeline. Provisions banning real-time biometric surveillance in public spaces and AI systems designed to manipulate human behavior still apply as scheduled. The extension is surgical, not a full retreat. For companies operating AI in Europe, the agreement is breathing room. For critics, it is confirmation that the EU began softening its most ambitious regulatory push almost before the ink dried. The question is whether a 16-month extension becomes a permanent one. Keywords: EU AI Act extension 2026, high-risk AI compliance deadline, European AI regulation, AI Act timeline delay