The exponential growth of artificial intelligence will require millions of dollars in investments for electrical infrastructure and data centers, according to Goldman Sachs.
The artificial intelligence (AI) revolution is transforming not only technology but also global energy consumption. According to a report by Goldman Sachs, data centers, which are already significant electricity consumers, will see their energy demand increase by 165% by 2030 due to the push from AI. This growth presents significant challenges for electrical infrastructure and necessitates multimillion-dollar investments to avoid bottlenecks in the tech sector.
Currently, data centers consume around 55 Gigawatts of energy globally. Of this total, 54% corresponds to cloud workloads, 32% to traditional business functions such as data storage or email, and the remaining 14% is allocated to AI. However, projections indicate that this landscape will change dramatically in the coming years.
An exponential jump in consumption
By 2027, energy demand from data centers is expected to reach 84 Gigawatts, with AI accounting for 27% of total consumption. This represents a 50% increase from current levels. Moreover, Goldman Sachs forecasts that by 2030, this figure will rise to 122 Gigawatts, with a 165% increase in demand associated with AI.
This explosive growth is driven by the race among major tech companies to develop increasingly powerful and complex AI models. In just 2024, AWS invested $75 billion in infrastructure, while Microsoft plans to spend $80 billion in 2025 and Meta will allocate $60 billion. These investments aim not only to expand the capacity of data centers but also to improve the energy efficiency of cooling systems, which consume between 35% and 40% of the total energy in data centers.
Challenges for the electrical grid
The rise of AI is also putting pressure on global electrical grids. Goldman Sachs estimates that $720 billion in investments will be needed to expand grid capacity and ensure stable supply through 2030. Without these improvements, the growth of the tech sector could be hindered by a lack of energy infrastructure.
This issue is especially urgent in Europe, where energy demand, after 15 years of decline, is rebounding due to the growth of data centers. According to the report, requests for electrical grid connections have multiplied in recent years, and energy demand on the continent is expected to increase by between 10% and 15% over the next decade.
The future of data centers
In the short term, Goldman Sachs foresees a tension between supply and demand in the data center sector, with occupancy rates potentially exceeding 95% by 2026. However, the situation should ease starting in 2027, thanks to the opening of new facilities and the development of more energy-efficient AI models.
The report also highlights the investment opportunities arising in this context, both for hyperscalers (large cloud infrastructure providers) as well as for asset managers, data center operators, and utility companies. The ability of these companies to adapt to the growing demand will be crucial for the future of the sector.
In summary, AI is not only transforming technology but also the global energy landscape. Its impact on data centers and electrical grids will require innovative investments and strategies to ensure sustainable growth in the coming years.