The European Union officially activated enforcement powers under the 2024 EU AI Act on August 2, 2026. The European Commission can now inspect frontier models, restrict market access, and fine providers up to 15 million euros or 3 percent of annual turnover, creating immediate compliance pressures for firms like Anthropic and OpenAI.
New EU Enforcement Powers Take Effect for General-Purpose AI
The European Union has shifted its regulatory approach from compliance guidance to active supervision. Under provisions that became fully applicable on August 2, 2026, as reported by cryptobriefing.com, the European Commission holds direct oversight authority over general-purpose artificial intelligence models. The enforcement activation follows a staggered rollout of the landmark 2024 legislation, closing out transition periods for advanced systems placed on the market after August 2025.
Regulators now possess the legal mandate to evaluate models before public release in the region, and order immediate operational halts if a system poses systemic risks. Non-compliant developers face severe penalties. According to Streamlinefeed, the European Commission can fine AI providers up to 15 million euros (approximately KES 2.1 billion) or 3 percent of their global annual turnover, whichever is higher.
Legal experts emphasize that the rules reach well beyond the borders of the member states. Elisabetta Righini, a partner at Sidley Austin, told CNBC that the regulatory powers apply to any organization offering a general-purpose AI model inside the bloc.
“A U.S. address does not put a lab outside the EU regulator’s reach,”
Elisabetta Righini, partner at Sidley Austin via CNBC
Righini also noted that non-European providers must appoint an authorized representative based in the EU to serve as the regulator’s point of contact. Furthermore, liability extends beyond substantive safety breaches; refusing information requests, providing misleading answers, or blocking model evaluations carries independent financial penalties.
Anthropic Shares Access to Mythos Amid Preemptive Compliance Push
Frontier artificial intelligence labs are altering their strategies to navigate the strict regulatory regime. Anthropic took the preemptive step of sharing access to its unreleased, advanced model codenamed “Mythos” with European Union officials, as Streamlinefeed detailed. The European Union had requested access to evaluate severe cybersecurity concerns associated with the model.
By permitting inspectors to probe the unreleased technology prior to public launch, Anthropic positions itself as a safety-first operator, establishing a cooperative stance that sets it apart from competitors. At the same time, OpenAI confirmed active communications with the regulatory body. Tom Gordon, vice president of EMEA policy at OpenAI, addressed the new oversight framework in a statement to CNBC.
“We’ve collaborated closely with the European Commission and the wider ecosystem on implementing the AI Act, including its Codes of Practice, and will continue working together to help Europe realise the benefits of the Intelligence Age,”
Tom Gordon, VP, EMEA policy at OpenAI via CNBC
Google also signaled its intent to meet regulatory expectations. A company spokesperson told CNBC that Google remains dedicated to fulfilling applicable rules while advancing European AI infrastructure and innovation.
Enterprise Deployment Ventures and Structural Market Bottlenecks
While managing regulatory compliance in Europe, leading artificial intelligence developers are simultaneously pursuing corporate acquisitions to streamline enterprise adoption. Joint ventures established by OpenAI and Anthropic are actively negotiating to acquire specialized services and consulting firms, according to reporting by channelnewsasia.com.
These initiatives address a central tension in the industry: deploying advanced machine learning models at scale requires labor-intensive, highly skilled engineers to tailor software to specific client data and workflows. OpenAI is raising approximately $4 billion through a vehicle named The Deployment Company with backing from 19 investors, including TPG, Bain Capital, and Brookfield Asset Management. Meanwhile, Anthropic is raising $1.5 billion from investors such as Blackstone, Goldman Sachs, and Hellman & Friedman.
Jon Gray, president and chief operating officer of Blackstone, explained the rationale behind the investment strategy in a statement published by channelnewsasia.com.
We believe it can help break down one of the most significant bottlenecks to enterprise AI adoption by expanding the number of highly skilled implementation partners,
Jon Gray, president and chief operating officer of Blackstone via Channel NewsAsia
Broader Geopolitical Pressures and Market Confidence
The activation of the European enforcement powers coincides with broader international frictions over technology policy and sovereignty. CNBC noted that the regulatory steps arrive amid ongoing sparring between Washington and Brussels over European tech sovereignty initiatives and penalties levied against American technology corporations, including a $1 billion fine issued against Google in July.
Despite the mounting regulatory pressures, prediction markets indicate strong investor confidence in major AI developers. According to data highlighted by cryptobriefing.com, Anthropic’s valuation is currently priced at an 87.5 percent probability in prediction markets to reach $1.25T by December 31, 2026. Observers continue to monitor ongoing communications between the European Commission and foundation model providers as enforcement actions proceed.
Worth a look