The expansion of artificial intelligence (AI) is redefining the balance between technological innovation and sustainability, placing energy consumption as one of the main challenges for the sector and for global technological development. In this context, data centers, as key infrastructures for the development of AI, currently represent between 1% and 1.5% of global electricity consumption, in a scenario of growing demand driven by the intensive use of digital services and the rapid development of advanced generative AI models.
Despite this growth, advances in energy efficiency have made it possible to moderate the impact of the exponential increase in data traffic, which has multiplied by 25 since 2010, while the number of internet users has more than doubled in the same period, reinforcing the sustainable development of the digital ecosystem.
In this context, MIOTI Tech & Business School, a leading business applied technology school in artificial intelligence and Data Science, defends the development of a more efficient and sustainable generative AI. The objective is to move towards models capable of optimizing the consumption of energy resources as a strategic axis of its development, in an environment of strong growth where the energy impact of this technology could exceed 3.6% of global electricity consumption in 2030, according to estimates by the International Energy Agency (IEA).
“We are facing a revolution unprecedented in its speed and scope, but we still cannot consider it sustainable in its current development,” says Fabiola Pérez, CEO of MIOTI. “The challenge is not to stop its adoption, but to make it more efficient and responsible in its development.”
Artificial Intelligence presents a double aspect in its development. On the one hand, it implies a high consumption of resources and, on the other, it offers great potential to promote sustainability in the business environment. Its ability to analyze large volumes of data allows organizations to measure their ESG impact, optimize production processes and improve energy efficiency in their operations, thus reinforcing their sustainable development. In this area, solutions such as Clarity AI show how Artificial Intelligence can contribute to the analysis, measurement and reporting of information linked to sustainability and impact, facilitating more precise decision making aligned with the responsible development of the business.
More efficient models for a more sustainable AI
MIOTI focuses on the development of smaller and specialized models, capable of offering high-value results with a lower computational cost. This trend responds to the need to adapt technology and its development to specific use cases, avoiding the unnecessary use of large-scale models in simple tasks. Therefore, several key lines are identified to move towards a more sustainable Generative Artificial Intelligence through the development of:
- Use of smaller and specialized models (Small Language Models): significantly reduce consumption compared to massive models and favor more efficient development.
- Algorithm optimization: techniques such as quantification allow the number of parameters to be reduced without losing performance, improving their technical development.
- New neural network architectures: recent advances are managing to reduce consumption in both training and use, driving the development of more sustainable systems.
- Responsible use of technology: simple tasks can be solved with less energy-intensive tools, promoting a more balanced development.
In parallel, technological infrastructure is also evolving as part of the development of the sector. Large technology companies are betting on integrating renewable energy into their data centers, as well as developing more flexible and decentralized infrastructure models, capable of taking advantage of energy surpluses in different locations. Along these lines, companies like Soluna are developing data centers located next to renewable energy sources, such as wind or solar farms, to transform surplus energy into computing capacity for intensive loads such as artificial intelligence.
At the same time, the debate on AI energy supply is opening the door to new alternatives in energy development, such as small modular reactors (SMR), a smaller-scale nuclear technology that some technology companies are beginning to consider as a way to guarantee stable, continuous and low-emission energy for their computing infrastructures and future development.
Large technology companies are betting on renewable energy in their data centers, and the development of more flexible and decentralized infrastructure models
“Generative Artificial Intelligence has entered a phase of maturity in which efficiency will be key to its development. Adoption is unstoppable and will continue to grow in all sectors, so the challenge is no longer just to develop more powerful models, but to use them more intelligently, design efficient systems from the beginning and measure their real impact in energy terms within their development,” says Fabiola Pérez, CEO and co-founder of MIOTI. “The future involves combining innovation, efficiency and responsibility in the development and use of technology.”
With growing global investment and accelerated adoption in the business environment, generative artificial intelligence is consolidated as a strategic technology for the development and competitiveness of organizations.
