This month, artificial intelligence is experiencing one of its most paradoxical moments since the beginning of the decade. On the one hand, financial markets continue to inject staggering sums into generative AI laboratories: Anthropic, one of the leading language model publishers, is now reportedly valued at around $965 billion, a figure that places it among the most highly valued technology companies in the world, just a stone's throw from the symbolic trillion-dollar threshold. Several financial sources even mention preparations for large-scale IPOs in the sector.
The employment paradox
On the other hand, this stock market narrative clashes with a harsher reality on the employment front. The technology sector reportedly cut over 200,000 jobs in recent months, a wave of cuts affecting both large corporations and startups, often justified by the increasing automation of tasks previously performed by employees. This tension between valuation euphoria and the contraction of tech employment fuels a fundamental debate: does AI create more wealth than it destroys, and for whom?
The AI Act's moment of truth
On the regulatory front, the deadline of August 2 is fast approaching. On this date, the European Artificial Intelligence Act, the AI Act, is set to activate its transparency obligations for general-purpose models, whether GPT, Claude, Gemini, or Mistral. From this deadline, the European AI Office could have initial leverage to sanction model providers who do not comply with documentation and transparency requirements. A Digital Omnibus intended to adjust certain provisions of the text still needs to be formally adopted by the European Council before this date.
Intellectual property war
The geopolitics of AI is also intensifying in terms of intellectual property and cybersecurity. Chinese laboratories have reportedly claimed to have identified security flaws in the code of certain Western models, while Anthropic has for several months accused players like Alibaba of practicing a form of « illegal distillation », i.e., training new models from the outputs of competing models without authorization. This legal and technical battle could set a significant precedent for intellectual property protection applied to AI models.
Computing credits and startup dependence
Another signal of the influence race between major laboratories: the massive distribution of computing credits to startups. Some startups reportedly receive the equivalent of several million dollars in computing benefits, a sum comparable to a classic seed round. This strategy allows AI giants to ensure lasting influence over the nascent ecosystem, making startups dependent on their infrastructure from their very first steps.
The ECB sets its conditions
The banking sector is not immune to this regulatory wave. An official from the European Central Bank has reportedly set a clear deadline for major European banks: to present, by the end of October, a plan for managing risks related to the use of advanced AI models. This requirement reflects a growing awareness among financial regulators, who now consider AI-related cybersecurity as a governance issue in its own right, and no longer as a mere technical matter relegated to IT departments.
The UN sounds the alarm
Finally, an independent scientific group mandated by the United Nations General Assembly recently sounded the alarm: the capabilities of AI systems are now progressing faster than the ability of states and the scientific community to understand and regulate them. Models are no longer limited to generating text or images, but are beginning to act autonomously, make decisions, and even bypass certain instructions given to them.
The editorial board's opinion
This July 2026 will remain a pivotal moment: when AI definitively ceased to be a mere technological promise to become a full-fledged economic, geopolitical, and social issue, with concrete repercussions on employment, finance, security, and regulation on a global scale.
Key takeaways
- $965 billion: Anthropic's valuation, nearing the $1 trillion club.
- 200,000+ tech jobs cut in recent months.
- August 2, 2026: AI Act transparency obligations come into force.
- End of October: ECB deadline for banks' AI risk management plans.
- UN warns of increasingly autonomous AIs, outpacing regulation.
