Ahmed Shaami
The European Union is moving to push back the toughest obligations of its landmark AI Act, delaying compliance deadlines for high-risk AI systems from August 2, 2026 to December 2, 2027, according to a report from law firm Holland and Knight and further detailed by Latham and Watkins. The change, part of a broader “Digital Omnibus” package, gives companies operating high-risk AI systems in Europe an additional 16 months to build out the conformity assessments, registration, risk management, and human oversight processes the law originally required by early August 2026.
What Was Supposed to Happen This Month
Under the AI Act’s original timeline, businesses running high-risk AI systems, covering areas like hiring, education, healthcare, finance, and biometric identification, needed to have those systems conformity-assessed, registered, and operating with full risk management and logging infrastructure in place by August 2, 2026. That deadline is now effectively suspended pending the Digital Omnibus legislation, which was expected to be finalized ahead of the original date.
Why Regulators Blinked
The delay reflects a gap between the law’s ambition and the state of industry readiness. As of April 2026, roughly 78 percent of organizations subject to the AI Act had not taken meaningful steps toward compliance, according to Holland and Knight’s analysis. Harmonized technical standards and practical implementation guidance needed to actually meet the law’s requirements were also unlikely to be published in time for the original deadline, leaving companies with little runway to adapt even if they wanted to move faster.
Not a Reason to Slow Down
Legal advisors are cautioning businesses against treating the delay as a reason to deprioritize AI governance work. Even with the new December 2027 timeline, companies should continue building out compliance and governance structures now, since the underlying technical standards and guidance remain in development and could still arrive with limited lead time before the revised deadline takes effect.
What It Means for Companies Building With AI
For beingguru’s audience of tech professionals, freelancers, and business owners building products or services on top of AI tools, the EU AI Act remains one of the most consequential regulatory frameworks shaping how AI can be deployed commercially, even with the extra time. Any product that touches hiring decisions, credit scoring, biometric verification, or other high-risk categories and serves EU users or customers should still be tracked closely, since the compliance bar has not been lowered, only pushed back on the calendar.
