Several tech innovations changed the way businesses operate, how products are designed, and how regulations govern the digital space in 2024. Three key areas stand out: the maturation of artificial intelligence, the energy consumption of digital infrastructures, and the arrival of European regulatory frameworks that alter the rules of the game for the entire sector.
Energy Consumption of Data Centers: The Hidden Cost of Artificial Intelligence
Have you noticed that every new generative AI application promises spectacular productivity gains? Behind these promises are servers, and these servers consume electricity, a lot of electricity.
The International Energy Agency estimates that data centers consumed about 415 TWh of electricity worldwide in 2024, which is around 1.5% of global demand. To put this in perspective, it is more than the annual consumption of many medium-sized European countries.
This figure, published in the IEA’s Energy and AI report in April 2025, places energy infrastructure at the center of the debate on technological innovations. Developing a high-performing AI model is no longer just a question of budget or talent, but also of electricity supply and server cooling.
Several industry players are seeking answers: relocating data centers to areas with decarbonized energy, liquid immersion cooling, optimizing queries to reduce server load. As reported by Scoopify, the tech news site, these issues are becoming as strategic as the development of the algorithms themselves.
European AI Act: What the Regulation Changes for Businesses
The European AI Act became law in 2024. It is the first legal framework in the world that classifies artificial intelligence systems by risk level and imposes proportionate obligations.
Why should this law catch your attention? Because its obligations are phased over time. The entry into force does not mean an immediate application of all the rules. Companies have different deadlines depending on the risk category of their AI system.
Specifically, systems considered to pose an unacceptable risk (social scoring, mass biometric surveillance in public spaces) are the first to be affected. High-risk systems, used for example in recruitment or medical diagnosis, follow a longer timeline.
What This Means for a SME or Startup
A company using an AI tool to sort applications will need to document how this tool works, assess its potential biases, and provide for a human oversight mechanism. This is not theoretical: compliance becomes a full-fledged budget item.
Companies developing AI applications for the European market should anticipate these requirements rather than discovering them at the time of application. The preparation window is open, but it is not infinite.
Cyber Resilience Act: Cybersecurity Required for Connected Products
The other major regulatory text of 2024 concerns the cybersecurity of connected objects. The Cyber Resilience Act establishes security requirements for all products containing digital elements sold in the European Union.
Until now, a manufacturer of connected cameras or smart thermostats had no legal obligation under European law regarding security updates or protection against known vulnerabilities. This regulatory ambiguity has facilitated the market introduction of products with minimal cybersecurity.
Timeline and Scope of the Text
The full application of the Cyber Resilience Act is scheduled for December 2027. By then, manufacturers must adapt their design, testing, and post-market monitoring processes.
The obligations cover several aspects:
- Integrate security from the product design phase, not as an added layer afterward
- Provide security updates throughout the expected lifespan of the product
- Report actively exploited vulnerabilities within a short time after their discovery
- Document the software components used to facilitate the identification of flaws
For consumers, this means that digital products purchased in Europe must offer a verifiable level of protection. For manufacturers, it is a complete overhaul of industrial practices.
Generative Artificial Intelligence in Business: Beyond the Hype
Generative AI dominated tech conversations in 2024. But beyond the impressive demonstrations, what has really changed in businesses?
Adoption has focused on specific tasks: writing reports, generating code, analyzing large documents, automated customer support. Productivity gains exist, but they heavily depend on the quality of internal data and team training.
The companies that have made the most of these tools are those that identified targeted use cases before deploying the technology. Those that adopted generative AI without a clear strategy often found themselves with high subscription costs and disappointing results.
The Issue of Proprietary Data
A technical point often overlooked: connecting an AI model to a company’s internal data raises questions of privacy and quality. A model trained on public data does not understand an organization’s specific processes. Adapting it requires time, skills, and a data governance framework.
Companies making progress in this area are investing in structured knowledge bases and in profiles capable of bridging business and technology.
Thus, 2024 has laid foundations that go beyond the mere adoption of tools. Between the energy pressure of digital infrastructures, the European legal framework for AI and cybersecurity, and the maturation of generative artificial intelligence use cases, companies’ technological choices are now inseparable from concrete regulatory and environmental constraints.
