Here’s what’s on our plate today:

• 🇪🇺 EU delays AI Act rollout under pressure from Big Tech.

• 📰 Figma’s India pivot, Google’s AI shift, and child safety tools.

• 🛠️ And this weekend, you can try: Canva Magic Studio, Gamma Decks, Yippity.

• 🗳️ Is the EU’s AI delay a smart strategy or self-sabotage?

Let’s dive in. No floaties needed…

The Laboratory

Can the EU’s regulatory delay help address its AI ambitions?

A durable regulatory system must be able to evolve with its environment. As societies change, economically, culturally, socially, or technologically, rules that once made sense can become outdated, incomplete, or misaligned with long-term goals. The strength of any framework lies in its capacity to adjust without losing its core purpose.

The U.S. Constitution offers a clear illustration of this principle. Since it came into force, it has been revised 27 times, reflecting lawmakers’ recognition that new realities demand new provisions. Yet revising rules is a double-edged exercise: amendments can modernise and correct, but they can just as easily dilute original safeguards or distort the spirit of the law.

In 2025, as global and economic powerhouses race to develop powerful AI systems, the European Union is poised to make a difficult decision.

After becoming the first major global power to formulate and begin implementing regulations to protect social and economic structures from powerful AI systems, the EU now appears prepared to slow its own regulatory rollout in the name of AI sovereignty.

But, scratch the surface, and it becomes clear that regulation is only one piece of a much broader structural challenge.

Europe’s AI Act meets resistance

The AI Act, the world's first comprehensive AI legislation, entered into force in August 2024 with a staggered implementation timeline. Yet by early 2025, key provisions were pushed well into 2026 amid criticism from industry and pressure from the EU’s largest trading partner.

Large corporations argued that regulatory inconsistencies could stall model development just as China accelerated its own capabilities.

In July 2025, to slow down the progress of regulations, business leaders, including senior figures from Airbus, BNP Paribas, Mercedes-Benz, and Philips, called for a two-year postponement, warning that "Europe's AI ambitions" were at risk due to unclear, overlapping and increasingly complex regulations.

U.S. influence adds weight to industry pushback

Washington amplified these concerns, and in April 2025, the US government's Mission to the EU reached out to the European Commission to push back against the AI code.

At the Paris AI Summit, U.S. Vice President J.D. Vance said that Europe risked strangling its own industry through excessive regulation.

Meanwhile, inside Europe, Mario Draghi’s 2024 competitiveness report similarly argued that premature regulatory burdens had historically pushed high-growth startups abroad, and recommended pausing the full suite of rules for high-risk models until impacts were better understood.

These pressures intersected with internal realities. While the EU faced pressure from the U.S. and big tech, deeper economic and structural concerns further pushed the commission to delay implementation.

The AI Act’s rollout was repeatedly hindered by missed deadlines and incomplete guidance.

Structural weaknesses beneath the delay

Since they were first announced, the AI Act's implementation has been plagued by delays that undermine its credibility. Guidance on prohibited AI practices was only released on February 4, 2025, days after the provisions came into effect.

This meant that organizations had very little time to consider the guidance before the regulation came into effect.

Similarly, regulations for general-purpose AI models were scheduled to take effect in August 2025, but the core frameworks were still unfinished. Part of the delay came from the Commission’s reluctance to miss deadlines at a time when the region is chasing a projected €3.4 trillion AI-driven economic boost by 2030.

The classification regime introduced further uncertainty: early assessments underestimated the number of systems likely to be considered high-risk, imposing heavier compliance burdens than anticipated. And while general-purpose AI rules were slated for August 2025, essential frameworks remained unfinished.

All these factors, when combined, were enough to push the EU to delay implementing its AI regulations.

Recently, an EU spokesperson told the Financial Times that some specific elements of the AI Act might face delays, though they stressed that the bloc is still fully committed to the law and its goals.

However, external pressure only tells part of the story. And, though policymakers frequently blame excessive regulation for slowing innovation, the EU’s challenge isn’t just overregulation; it’s underinvestment.

Europe’s investment and infrastructure gap

In 2024, AI-focused venture investment in Europe lagged far behind the levels seen in the United States and China, and the pattern was largely unchanged from 2023. This shortfall in capital has had a visible impact: by 2025, the U.S. was turning out multiple cutting-edge foundation models, China was narrowing the gap, and Europe had produced only a small handful in comparison.

The imbalance is even greater when viewed through the lens of infrastructure.

According to a report from McKinsey, roughly 70% of Europe’s digital services run on U.S. hyperscalers, and the continent remains reliant on advanced chips designed abroad and fabricated in Asia, with local semiconductor production representing less than a tenth of global capacity.

These weaknesses reflect an underdeveloped innovation ecosystem, one where the flywheel of research funding, venture capital, risk-tolerant institutional investment, and pension-driven long-term capital simply has not reached the scale needed to compete.

So, while regulators are willing to delay the implementation of regulations, the lack of a cohesive plan for the development of AI infrastructure, models, and startups that make it profitable may be hampering progress.

And though policymakers and business leaders have responded with high-profile commitments and ambitious targets, results have lagged behind the rhetoric.

A proposed €150-billion campaign to boost European AI startups, along with major national investment pledges such as France’s €109-billion data-center plan, signal serious political intent.

Yet implementation remains slow and fragmented. All this has resulted in an environment where Europe articulates big aspirations for AI leadership but struggles to build the financial, infrastructural, and industrial foundations needed to make those ambitions real.

The cost of hesitation

So, while the EU may be looking to delay regulatory oversight, it might prove to be counterproductive in the long run.

Regulatory clarity provided by the AI Act could help facilitate investment decisions and product development strategies. Some view regulations as a strategic advantage.

The regulation's extraterritorial reach means that American, Asian, and other international AI companies must comply with EU standards when their systems affect European users, creating a Brussels Effect similar to GDPR's global impact on data privacy practices.

The real test of European AI sovereignty

The premise that halting AI rules will enable Europe to catch up to the U.S. or China is unproven. And could have the opposite impact. Europe’s reliance on American hyperscalers could further increase due to a lack of regulatory oversight.

The real question then is whether Europe can build the financing ecosystems, cloud infrastructure, and semiconductor supply chains necessary to compete. These are challenges that regulatory simplification alone cannot address. The trajectory of AI development will not be determined by regulation or replication, but by the entrepreneurial spirit that drives innovation forward.

The EU’s constitution, much like the U.S. Constitution, has the ability to adapt to changing times. And while Europe was early to recognise the need for guardrails on powerful AI, delaying implementation may, ironically, weaken both its regulatory authority and its long-term competitiveness.

TL;DR

• The EU is delaying parts of its AI Act amid industry backlash, U.S. pressure, and internal disarray.

• Implementation lags due to vague guidance, underestimated risk classification, and unfinished frameworks.

• Europe’s deeper problem? Chronic underinvestment in AI infrastructure, chips, and startups.

• Regulatory delays won’t solve the core issue: the EU needs more builders, not just better laws.

Headlines You Actually Need

• Figma’s India Play: The design giant pivots to India for non-design expansion.

• Google’s Private AI Push: New tools let users train models without cloud exposure.

• AI vs. Abuse Content: Big Tech teams up to test AI’s ability to detect child abuse imagery.

Friday Poll

Weekend To-Do

• Canva Magic Studio: Try Canva’s new AI-powered design suite — text-to-image, video editing, and Magic Write are now bundled into a single interface.

• Gamma AI Decks: Instantly create sleek, pitch-worthy slide decks using just a prompt. Great for summarizing ideas, reports, or turning Roko pieces into visuals.

• Yippity: Paste any link, and this tool turns it into a quiz to help you retain what you read — ideal for newsletters, articles, even PDFs.

Rate This Edition