The year 2024 has seen several technological mutations crystallize, the effects of which go beyond the usual marketing announcements. Between the enforcement of the first binding legal framework on artificial intelligence in Europe, the growing pressure on water resources from data centers, and the maturation of generative AI in business processes, the technological landscape has been reconfigured on more concrete bases than expected.
Water Footprint of Data Centers: The Invisible Constraint of AI
Prospective analyses on technological trends for 2024 consistently mention the electricity consumption of digital infrastructures. However, they overlook an equally crucial factor: water.
A report from the Lawrence Berkeley National Laboratory published in 2024 estimates that American data centers directly consumed approximately 17 billion gallons of water in 2023. The indirect consumption, related to the electricity production powering these facilities, is significantly higher.
This data changes the interpretation of expansion projects. The choice of site, local climate, and the cooling technology selected determine the actual impact of a data center on the region’s water resources. In arid regions or those under water stress, local opposition to new projects is intensifying, complicating deployment timelines. Several technological watch reports published on www.kalinews.net document this tension between digital expansion and environmental limits.
The available data does not yet allow for precise measurement of the share attributable solely to generative AI workloads in this consumption. Field reports vary on this point depending on the cooling architectures employed.
AI Act: The European Regulation Redefining AI System Governance
While most 2024 reports discuss the ethics of artificial intelligence in general terms, a specific legal fact has transformed the framework applicable to European companies. The regulation (EU) 2024/1689 came into force on August 1, 2024.
This text introduces a classification of AI systems by risk levels. Four categories structure the framework:
- Prohibited systems, whose practices are deemed incompatible with fundamental rights (generalized social scoring, subliminal manipulation).
- High-risk systems, subject to strict compliance obligations before being placed on the market (automated recruitment, credit scoring, biometric identification).
- General-purpose models (foundation models), for which transparency and technical documentation requirements apply.
- Limited-risk systems, subject only to information obligations towards users.
For technology companies as well as sectors deploying AI tools (hospitality, services, industry), this regulatory shift implies a task of inventory and compliance that was not on the roadmaps a year earlier. The AI Act transforms AI governance into a legal obligation, rather than a mere voluntary commitment.
Generative AI in Business: Beyond the Experimentation Phase
The year 2024 has marked an intermediate phase where technology has moved from proof of concept to integration into operational processes.
Differentiation relies on three levers: pre-trained models accessible via API, cloud computing, and open source. Access to generative AI no longer requires internal research skills. Business teams without data science expertise are deploying writing assistants, document summarization tools, or customer recommendation systems.
This accessibility creates a new problem: the proliferation of unsupervised uses. Without structured safeguards to monitor models, protect data, and manage biases, productivity gains can come with uncontrolled automated decisions, directly impacting clients and sensitive data.
Documented Limitations of GenAI in Production
The hallucinations of language models remain an identified barrier in critical environments. Systems generate plausible but factually incorrect responses, necessitating human verification in any decision-making chain. The cost of this supervision reduces some of the expected savings.
The real cost/benefit ratio of generative AI depends on the sector and use case. Applications for generating marketing content or first-level customer support show measurable gains. In contrast, deployments in regulated fields (healthcare, finance, hospitality) require layers of validation that complicate integration.
Cybersecurity and Digital Resilience: The Cyber Resilience Act in the Background
The European Cyber Resilience Act complements the regulatory landscape of 2024. This text imposes security obligations on manufacturers of connected products throughout the entire lifecycle, with requirements for incident notification and security updates.
For companies, this trend implies that security becomes a design criterion rather than a post-deployment fix. IoT systems deployed in industry, connected hotels, or customer services must integrate resilience from the initial architecture.
Continuous Threat Exposure Management (CTEM) is gaining adoption alongside these legal requirements. The approach involves continuously identifying attack surfaces rather than conducting one-off audits.
Sustainable Technologies: A Market Criterion, Not Just Communication
Regulatory pressure and rising energy costs converge to make sustainability an operational parameter in technological choices. Companies investing in cloud infrastructures now evaluate the energy and water efficiency of suppliers alongside technical performance.
The choice of a cloud provider now incorporates measurable environmental criteria. Certifications, commitments on energy sources, and transparency on water consumption are becoming differentiating factors in tenders, particularly in the hospitality sector and business services.
The technological landscape of 2024 stands out from previous years by the prominence of external constraints: regulation, natural resources, operating costs. The most structuring trends are not those that promise the most, but those that compel organizations to rethink their practices under the threat of non-compliance or operational overcosts.
