This week, the European Union agreed to delay major parts of the AI Act, pushing high-risk AI compliance requirements to late 2027 while simplifying obligations for businesses and expanding flexibility for SMEs and mid-sized companies. Reuters described the agreement as a softened version of the original framework after pressure from governments and industry groups concerned that Europe risked slowing innovation while the US and China accelerated ahead.
At the same time, the EU introduced stricter bans targeting AI-generated non-consensual intimate imagery, reinforcing that Europe still plans to maintain tighter safeguards in specific areas. The European Commission stated that the updated agreement is designed to support innovation while maintaining safety standards.
The agreement also reflects a wider shift happening across Europe’s AI market. Alongside the AI Act discussions, governments and companies are increasing investments into compute infrastructure, semiconductors, cloud systems and AI deployment capacity.
Europe’s AI Act Was Supposed To Set The Global Standard. Now Europe Is Adjusting It.
According to Reuters, negotiations around the AI Act became increasingly difficult after several member states raised concerns around operational complexity, overlapping regulations and the impact on Europe’s competitiveness.
The updated agreement delays high-risk AI compliance timelines until December 2027, while AI systems embedded into products such as machinery receive longer transition periods. The Council of the European Union also confirmed that the revised framework expands regulatory sandboxes and simplifies certain obligations for smaller companies.
Industry groups including DigitalEurope and the Computer & Communications Industry Association supported the revisions, arguing that Europe needed a more practical framework for companies building AI products locally. Consumer organisations, however, warned that delaying parts of the AI Act could weaken protections and reduce accountability.
The changes suggest that Europe is trying to balance regulation with the need to scale its AI ecosystem more quickly.
Europe’s Startups Just Got More Time To Scale
The revised timelines are expected to have a direct impact on AI startups across sectors including healthcare, robotics, fintech, industrial AI, manufacturing, mobility and enterprise software.
Many startups had raised concerns that the original AI Act timelines would increase compliance costs too early, especially for smaller companies still developing products and raising capital.
The updated agreement gives startups additional time before stricter compliance obligations fully apply. According to the European Commission, the revised framework also expands access to AI testing environments and support programmes for SMEs and mid-sized companies.
This is particularly relevant for Europe’s deeptech ecosystem, where development cycles are longer and infrastructure costs are significantly higher than traditional software businesses.
Investment activity is already reflecting that shift.
Reuters reported earlier this year that Dutch AI semiconductor startup Axelera AI raised an additional $250 million to expand AI chip production in Europe. German startup Peak Quantum also secured new funding to expand locally manufactured quantum hardware systems.
Europe’s Focus Is Expanding Beyond AI Models
Alongside the AI Act discussions, Europe is accelerating investments into compute infrastructure, semiconductors, cloud systems and AI capacity.
According to the Council of the European Union, Europe is moving forward with plans for AI “gigafactories” designed to strengthen regional compute power and reduce dependence on foreign infrastructure providers.
Europe’s first exascale supercomputer, JUPITER in Germany, recently crossed one exaflop of performance using more than 24,000 NVIDIA GH200 GPUs, placing Europe among the few regions globally operating at that scale.
Bloomberg also reported renewed momentum around a possible “Chips Act II” initiative that could allow the European Commission to directly invest in semiconductor fabrication facilities across the region.
At the same time, Reuters reported that EU regulators are increasing scrutiny around cloud infrastructure and AI services, particularly involving major US hyperscalers such as AWS and Microsoft Azure.
Amazon recently launched a Europe-based sovereign cloud service designed to address concerns around data control and infrastructure dependency inside Europe. According to Reuters, AWS Germany CTO Michael Hanisch said the infrastructure was designed to continue operating even under scenarios involving internet fragmentation or US export restrictions.
Enterprises Are Watching The Changes Closely
The updated AI timelines are also important for enterprises across manufacturing, healthcare, automotive, banking, energy, logistics, and industrial sectors.
Many large organisations spent the past year slowing or carefully managing AI deployment plans due to uncertainty around regulation, governance, and compliance obligations.
The revised framework now gives enterprises clearer timelines while reducing immediate compliance pressure. At the same time, infrastructure and cloud sovereignty are becoming larger priorities for companies operating in regulated industries.
Where This Converges: GITEX AI EUROPE
The topics shaping Europe’s AI strategy this week are also becoming central discussion areas across GITEX AI EUROPE. As Europe adjusts its approach to AI regulation, enterprises, startups, policymakers, investors and cloud providers are increasingly focused on competitiveness, compute access, enterprise adoption and reducing dependency on external infrastructure.
These discussions are becoming increasingly relevant across, where AI deployment is now moving into large-scale implementation. At GITEX AI EUROPE, many of these conversations will move beyond policy and into industry execution, bringing together the companies, investors, infrastructure providers and government leaders shaping Europe’s next phase of AI growth.