The European regulation on artificial intelligence (AI Act) came into effect on August 1, 2024. This date is a game changer for the entire IT sector: AI is no longer just a productivity lever; it becomes a matter of regulatory compliance with obligations graded according to the risk level of the deployed systems.
AI Act and NIS2 Directive: the regulatory framework redefining IT in 2024
The entry into force of the AI Act requires organizations to classify their AI systems by risk categories. The obligations do not apply uniformly: the application is progressive according to the actors and categories of systems. Providers of foundation models, integrators, and end users are not subject to the same deadlines or documentation requirements.
At the same time, the NIS2 directive set October 17, 2024 as the deadline for transposition into national law. It significantly broadens the scope of entities subject to cybersecurity obligations and strengthens the requirements for managing supply chain-related risks.
These two texts converge on one point: IT governance is no longer just an internal technical issue; it is an organizational responsibility challenge. CIOs who still treat compliance as a side project accumulate measurable legal risk. To keep up with these developments, IT information on Geeknology regularly covers these regulatory topics applied to the sector.
Generative AI in production: beyond the prototype
Generative language models have moved beyond the experimentation phase in many organizations in 2024. Their integration now affects marketing content creation, legal assistance, and search engine optimization. The question is no longer whether generative AI works, but how to govern it at scale.
The real bottleneck lies in the quality of training data and the management of biases. A model deployed without auditing its outputs exposes the organization to reputational risks and, now, regulatory risks under the AI Act.
Energy consumption of AI data centers
The International Energy Agency dedicated a specific analysis in 2024 to the electricity consumed by AI-related data centers. The rise of generative AI has an energy cost that constitutes a significant IT challenge. Queries on large models consume significantly more energy than a standard search.
This reality requires cloud architects to rethink infrastructure sizing. Green IT is no longer just about recycling hardware: it now incorporates algorithmic sobriety as a design parameter.
Cybersecurity in 2024: the supply chain as an attack surface
The NIS2 directive explicitly targets the security of the digital supply chain. Bounce-back attacks via third-party providers have become a major vector. We observe that the requirements of NIS2 compel clients to audit their IT suppliers with an unprecedented level of rigor.
The obligations focus on several concrete areas:
- The establishment of documented and auditable cyber risk management policies covering all critical information systems
- The formal assessment of the security posture of subcontractors and providers with access to the entity’s networks
- The mandatory notification of significant incidents within specified timeframes, with a rapid initial report followed by a detailed report
Cybersecurity is becoming a criterion for selecting suppliers, alongside price and availability. IT contracts increasingly incorporate NIS2 compliance clauses as prerequisites.
Sustainable technologies and green cloud computing: constraint or competitive advantage
The concept of sustainable technology encompasses two dimensions: solutions designed to address environmental challenges, and technologies produced through ecologically responsible processes. In 2024, this dual requirement directly applies to cloud infrastructures.
Hyperscalers are committing to carbon neutrality, but the actual balance depends on the energy mix of the regions where the data centers are located. A workload hosted in a center powered by renewable energy does not have the same impact as an identical workload in a high carbon intensity area.
Circular economy applied to IT hardware
Green IT also extends to the lifecycle of hardware. The so-called Green 4 IT approach aims to optimize equipment based on its entire lifespan, from manufacturing to recycling. For corporate IT parks, this translates into policies for refurbishment and extending renewal cycles.
The manufacturing of terminals (computers, smartphones) accounts for the majority of their carbon footprint. Extending the usage duration of a workstation by one or two years significantly reduces its impact, often more than software optimization.
Quantum computing and blockchain: real maturity in 2024
Quantum computing is gradually moving out of the experimental stage, but its adoption remains confined to very specific use cases: molecular simulation, combinatorial optimization, cryptanalysis. No quantum computer today replaces a traditional server in production.
Blockchain follows a similar trajectory. Industrial applications are progressing (traceability, smart contracts), while public adoption stagnates. The projects that succeed in 2024 are those that solve a specific business problem, not those that deploy the technology for its own sake.
- Quantum computing interests the pharmaceutical and financial sectors for simulation calculations inaccessible to classical architectures
- Blockchain finds its relevance in certifying supply chains and tokenizing real assets
- Immersive experiences (augmented reality, virtual reality) are progressing in professional training and industrial maintenance, but the consumer market is still waiting for a catalyst
The IT trends of 2024 share a common trait: technology alone is no longer enough. The AI Act, NIS2, energy constraints, and market maturity require a combination of technical capability and governance framework. Organizations that integrate this dual perspective into their IT roadmap gain a concrete advantage over those that merely adopt tools without measuring the implications.
