The year 2024 has reshuffled the priorities of the technology sector around three axes: the regulation of artificial intelligence, the energy constraints of digital infrastructures, and the material sustainability imposed by the European framework. These three dynamics are concretely changing how companies deploy their tools and how users interact with their devices.
Energy constraints of data centers and AI deployment
Before discussing increasingly efficient AI models, it is essential to understand what physically limits their deployment. Data centers, which host the training and inference of models, consumed approximately 415 TWh of electricity worldwide in 2024 according to the International Energy Agency.
This figure places electrical availability as a decisive factor for the establishment of new infrastructures. Access to reliable and sufficient power now takes precedence over network connectivity in site selection. A data center may have the fastest fiber on the market, but without stable electricity, it remains unusable for training large models.
This physical reality hinders the large-scale democratization of generative AI. Cloud providers are readjusting their expansion plans based on available electrical networks, prompting some regions to invest heavily in dedicated energy production. To keep up with these developments, you can check Info 11 for tech, which regularly covers industry news.
European AI Act: a progressive obligation timeline
The AI Act, the European regulation on artificial intelligence, came into effect in August 2024. Its uniqueness lies in its phased implementation: some obligations began as early as 2025, while general enforcement is expected in 2026.
This gap between the entry into force and actual implementation creates a transition period during which companies must prepare for compliance. The regulation notably requires a classification of AI systems by risk level, with each category leading to different requirements.
Energy documentation of AI models
A less discussed aspect of the AI Act concerns the documentation of the energy consumption of general-purpose models. Providers must retain technical information including known or estimated consumption during development. When direct measurement is not available, the regulation allows for estimation based on the computing resources used.
This obligation directly connects AI regulation to the energy constraints mentioned earlier. It encourages developers to integrate energy efficiency as a design criterion, not just as an operational cost variable.
Smartphone repairability: sustainability imposed by European regulation
Hardware innovation in 2024 is no longer measured solely by processor power or screen quality. European regulation has introduced concrete repairability requirements for smartphones, a shift that alters manufacturers’ design criteria.
These rules require manufacturers to make certain spare parts accessible and to provide repair information. The goal is to extend the lifespan of devices rather than encourage their systematic replacement.
- Manufacturers must guarantee the availability of spare parts for a defined period after the product is launched
- The technical information necessary for repairs must be accessible to independent professional repairers
- The repairability index becomes a commercial differentiation criterion, not just a regulatory constraint
This framework transforms sustainability into a selling point. Brands that anticipate these requirements gain a competitive advantage among consumers who are increasingly attentive to the lifecycle of their devices.
Generative AI in business: from experimentation to operational integration
Generative AI has dominated technological discussions since late 2022, but 2024 marks a phase shift. Companies are moving from experimentation to integration into their business processes.
This scaling raises questions that the initial phase of enthusiasm had sidelined:
- The governance of the data used to train the models, with issues of privacy and intellectual property
- The actual cost of the ongoing operation of generative AI tools, which far exceeds the advertised subscription price
- The reliability of the results produced, which requires human verification processes in critical professional uses
Trust and security management of AI
The notion of trust, risk, and security management of AI (AI TRiSM) is emerging as a structuring framework. This concept encompasses the continuous monitoring of systems, the ethical principles applied to algorithms, and the associated security protocols.
Organizations deploying generative AI tools without this type of framework expose themselves to reputational and legal risks, especially with the gradual implementation of the AI Act.
Sustainable technologies and responsible computing
The convergence between technological innovation and environmental constraints is no longer a communication topic but a design parameter. Energy efficiency now conditions software architecture choices as much as raw performance.
Industrial cloud platforms integrate carbon footprint measurement tools directly into their dashboards. This transparency allows companies to quantify the environmental impact of their technological choices and adjust them according to measurable objectives.
The technological trends of 2024 share a common thread: raw performance is no longer sufficient. European regulation, energy constraints, and material sustainability requirements are redefining what viable innovation means. Organizations that integrate these parameters from the design phase of their technological projects position themselves better than those that treat them as late administrative obligations.
