The debate on artificial intelligence (AI) often focuses on its social, cultural, and educational effects, but there is an urgent issue that deserves attention: the geopolitical and environmental impact that the expansion of AI generates in Global South countries, such as Argentina, which are rich in natural resources and vast territories. Beyond the so-called "AI bubble," concerns are growing about the energy consumption of these systems. According to Goldman Sachs, the five major tech companies will invest $736 million between 2025 and 2026, and the International Energy Agency estimates that by 2030, they will consume nearly 20% of global energy demand. The announcement of a potential OpenAI data center in Patagonia last year raised alarms about the loss of digital sovereignty and the associated environmental costs. This investment is part of the controversial Incentive Regime for Large Investments (RIGI), promoted by the Argentine government, which limits the state's ability to demand environmental, fiscal, and local production standards. Critics warn that this could facilitate the transfer of national resources to foreign interests. Meanwhile, Congress is debating a reform to the Glacier Law that could allow currently prohibited activities, such as mining and large projects in protected areas, which would represent a serious environmental setback. Yas García, director of AI and data at the Open Knowledge Foundation and Jump Education, points out the opacity of public-private negotiations: "They tend to be non-transparent, mixing land use, energy contracts, and tax benefits under commercial confidentiality." Regarding the Stargate Argentina project, promoted by Sur Energy and OpenAI, she warns that job creation will be limited and mainly during the construction phase. "The question is what the agreements really stipulate: land, energy, who finances the works?" she questions. In a context where major tech companies are closing circular agreements—OpenAI pays Oracle for computing capacity, Oracle pays Nvidia for chips, and Nvidia invests in OpenAI—the construction of data centers in remote locations is no longer unusual. Milagros Miceli, a sociologist and doctor in computer engineering at DAIR Institute and Weizenbaum Institute, highlights that data centers are key to attracting investors by offering tangible infrastructure that supports the promise of AI. However, she warns that the industry prioritizes growth and mass data consumption over more sustainable and efficient models. In Argentina, the trend towards deregulation favors a lack of controls and transparency. The Minister of Deregulation and Transformation, Federico Sturzenegger, has stated that his goal is to avoid any regulation on AI, raising doubts about the protection of national interests. Beyond environmental risks, there is a danger that Argentina could become a mere "depot" for data centers, losing autonomy and value in the global AI chain. The Argentine Node of Artificial Intelligence (Nadia) warns about the importance of negotiating conditions that ensure control, transparency, and local benefits. The key question is: what does Argentina gain from these investments? Digital sovereignty does not mean rejecting investments but establishing clear rules that ensure returns and protection of rights. García suggests that national companies like YPF and Arsat could lead a sovereignty strategy, but emphasizes that a state policy is required. "We need public infrastructure, a renewable and flexible energy system, and capabilities in universities, science, and the productive sector to capture value," García concludes. "If we only receive machines without developing our own capabilities, we will be a computing territory, not technology producers." Source: elpais.com