New technologies no longer refer solely to software or applications. They now encompass physical infrastructures (data centers, semiconductors, electrical networks) and the regulatory frameworks that govern their deployment. Understanding how these technologies are transforming the global economy requires looking beyond artificial intelligence models to examine what makes them possible and what holds them back.
Infrastructures and Semiconductors: The Material Foundation of the Tech Economy
Media attention focuses on AI models, but technological growth is hindered by physical constraints. According to the McKinsey Global Institute, the advancement of artificial intelligence depends less on the availability of algorithms than on access to data centers, electricity, cooling, and semiconductors.
Electronic chips, particularly advanced semiconductors, are at the center of a competition among countries. The United States, China, and India are heavily investing in local production capabilities. This semiconductor race is reshaping global supply chains and creating new geopolitical dependencies.
An article detailing the role of technologies on Claravox reminds us that these infrastructure issues directly condition the innovation capacity of companies, regardless of their sector.
The bottleneck is not limited to chips. Engineers specialized in semiconductor design and data center management are a rare resource. Training these profiles takes several years, which slows the pace of expansion even when funding is available.
Energy and Data Centers: The Tension Between Digital Growth and Decarbonization
The rise of data centers is generating increasing pressure on electrical networks. According to S&P Global, this demand may prolong reliance on gas production in certain regions, creating a direct tension between digital expansion and decarbonization goals.
This point remains underexplored in traditional analyses of the digital economy. Local opposition to the establishment of new data centers is intensifying in several European countries. Communities are measuring the gap between the jobs created (often few once construction is complete) and the water and electricity consumption generated.
For tech companies, this energy constraint is becoming a strategic location factor. Countries with abundant and decarbonized electricity (hydropower, nuclear) are attracting more digital infrastructure projects. Access to energy now conditions the technological competitiveness of a territory.
From Generative AI to Agentic AI: What Changes for Businesses
Generative artificial intelligence (capable of producing text, images, or code) has dominated discussions for the past few years. A more recent evolution deserves to be distinguished: the shift to agentic AI.
This term refers to systems capable of chaining multiple actions autonomously, managing processes, and assisting software development without human intervention at every step. The difference with generative AI lies in the execution capability: agentic AI does not just propose; it acts.
What This Means in Practice
For businesses, this transition alters the nature of expected productivity gains. Generative AI allowed for the acceleration of content or analysis production. Agentic AI promises to automate entire sequences of tasks:
- Management of logistical flows with real-time adjustments without intermediate human validation
- Assisted software development, where the agent writes, tests, and corrects code over multiple iterations
- Management of marketing campaigns with dynamic budget allocation across channels
This evolution raises questions of internal governance. Delegating decisions to a software agent requires precisely defining its limits of action and the cases where human intervention remains necessary.
Regulation of Artificial Intelligence: The AI Act Enters Operational Phase
The European Union has adopted Regulation 2024/1689, known as the AI Act. This text classifies AI systems by risk level and imposes graduated obligations on companies deploying them.
Unlike the principle statements that preceded it, the AI Act enters a phase of concrete application with a binding timeline. Companies must now document their systems, assess risks, and ensure transparency towards users.
Consequences for Economic Actors
Obligations vary according to the risk category of the system. High-risk applications (automated recruitment, credit scoring, biometric surveillance) are subject to strict audit and traceability requirements.
- Companies using large language models must assess the systemic risks of their tools
- Deployers of AI in health or justice are required to maintain permanent human oversight
- Providers of generative AI systems must clearly indicate that the content is machine-generated
For countries outside the European Union, the AI Act creates a regulatory ripple effect. Companies exporting to the European market will need to comply with these rules, prompting other jurisdictions to harmonize their own frameworks.
Innovation and Development: Contrasting Regional Dynamics
Technological growth is not evenly distributed. Countries with an ecosystem combining venture capital, technical talent, and reliable infrastructure capture the majority of the value created by digital innovation.
India, for example, is betting on the massive training of engineers and local semiconductor manufacturing to reduce its dependence on imports. The United States maintains an edge in fundamental research in artificial intelligence, while Europe seeks to compensate for its industrial lag with an attractive regulatory framework and targeted investments.
Developing countries remain largely consumers of technologies produced elsewhere. The United Nations highlights that digital technologies have reached nearly half of the population in developing countries over two decades, but access does not guarantee the ability to produce or capture the associated economic value.
The gap between technology-producing and technology-consuming countries is likely to widen if investments in infrastructure and training do not keep pace. The next decade will be marked by each economy’s ability to transform access to digital tools into real productivity gains, rather than merely surface adoption.
