August 2, 2026: The Day Europe’s AI Rules Become Real

Mark the date. On August 2, 2026, the EU AI Act becomes fully enforceable.

What has been a compliance timeline on paper becomes law in practice — with fines of up to €35 million, or a percentage of global annual turnover, for companies that fail to meet its requirements.
Every GAFAM company will be affected. The Act applies not only to European businesses but to any company whose AI systems reach EU users — regardless of where servers are located. Google, Apple, Meta, Amazon and Microsoft all qualify.
The high-risk categories are where the real pressure will be felt: AI systems used in hiring, healthcare, law enforcement, border control and credit assessment face the strictest obligations — mandatory impact assessments, human oversight requirements, and detailed technical documentation.

There is uncertainty about timing.

The European Commission has proposed adjustments to the high-risk deadline, partly due to delays in developing the technical standards companies need to demonstrate compliance. Whether August 2026 holds or slips into 2027 depends on political negotiations now underway in the European Parliament and Council.
What is not uncertain: the direction. Europe has decided that AI must be accountable, transparent and subject to democratic oversight. Washington has decided the opposite — deregulation, speed, and military application are the Trump administration's AI priorities.
These two visions are on a collision course. gafam.ai will be watching.

This is the story that defines our publication. The five most powerful technology companies in the world, operating between two irreconcilable regulatory philosophies. Welcome to the European eye on Big Tech and AI.