Investing in thematic funds: Is AI positive for the climate?

When a single announcement leads to temporary book losses of nearly USD 1 trillion, it certainly captures the attention of the investment community. This happened last January when DeepSeek unveiled its latest AI-equipped application, R1.

The Chinese startup revealed that it had developed the AI model using innovative algorithms in a much shorter time and at a fraction of the cost of its major US competitors. These competitors suddenly found their leading positions in AI threatened. What also alarmed investors and caused the stock prices of power suppliers to fall: The DeepSeek model apparently consumes much less electricity – up to eleven times less than a model from the American tech giant Meta, according to a study published in the scientific journal "Nature".

No AI Without Energy?

The DeepSeek shock also brought to the forefront a topic that has long been discussed among sustainability-focused investors – namely the significant dependency of future technology on energy supply. "No energy, no AI," is the pointed opinion of the International Energy Agency (IEA), which has dedicated a new study to this issue. The figures presented by IEA experts are striking: A typical data center, which seeks to draw intelligent conclusions from vast amounts of data, consumes as much electricity as 100,000 households. Combined, such centers consumed 415 terawatt-hours (TWh) of electricity in 2024 – roughly equivalent to Germany's energy consumption (see graphic below).

AI Data Centers Drive Electricity Demand (in TWh)

Can Renewable Energy Satisfy AI's Hunger?

The agency forecasts that this will increase significantly in the future. The energy consumption of data centers could double by 2030 and even triple by 2035. According to the study, AI is the largest energy consumer, despite less "energy-hungry" models like DeepSeek's. This leads to developments that investors interested in the themes of decarbonization and climate should be aware of, both positively and negatively.

This is evident in the energy sources that power data centers worldwide. Already today, renewable and low-carbon energy sources play an important role. According to IEA calculations, they can already meet half of the global electricity demand growth from data centers. However, electricity generation from natural gas and coal also plays a central role. Currently, coal and gas-fired power generation account for the other half of data centers' energy consumption. This will not change significantly by 2030.

Energy and Emissions: What Is AI's Climate Impact?

Accordingly, the electricity consumption of AI data centers raises concerns. Will the much-praised future technology become a climate killer due to its energy hunger leading to a rapid increase in greenhouse gas emissions?

According to IEA forecasts, there is some truth to this, at least at first glance. Along with road and air traffic, data centers are one of the few areas where the study sees direct and indirect CO₂ emissions continuing to rise. Emissions caused by the electricity demand of data centers could climb to around 475 megatons (Mt) by 2035 (see graphic below). However, AI could also significantly contribute to limiting greenhouse gas emissions and advancing the decarbonization of the economy, as we recently noted regarding the importance of digital environmental solutions.

AI applications can be used in various sectors, not least in energy supply. According to IEA forecasts, this could result in a positive net effect. The broader use of AI could reduce global CO₂ emissions by 1,400 Mt within ten years. This would be three times the additional climate burden caused by the expansion of data centers.

Positive Net Effect of AI on Energy Consumption and Emissions

Although the IEA expert team warns that AI is not a "silver bullet" for solving the climate crisis, the expected efficiency gains from AI open up exciting perspectives for investors interested in the sustainable investment theme of climate and decarbonization. The following directions should be considered:

• Optimization with AI: Energy supply is becoming increasingly digitalized, more interconnected, and simultaneously decentralized. This results in higher complexity, which calls for the use of AI. For example, since 2024, the Japanese industrial conglomerate Hitachi Energy has been selling an AI-based model for forecasting electricity demand, electricity generation from renewable energies, and electricity prices. According to the company, the model is 20% more accurate than conventional models. AI can also support so-called load shifting: Charging electric vehicles and heating or cooling rooms can be shifted to times of day when electricity is cheap and plentiful. According to the IEA, this could reduce peak electricity demand in industrialized countries by around 15%.
• Innovation through AI: AI can play an important role as a catalyst for research and development around sustainability. Consider the healthcare sector with drug development: In 2024, the Nobel Prize in Chemistry was awarded to John Jumper and Demis Hassabis for their AI model "Alpha" for predicting protein structures. Their AI model is also used in material research for new batteries. Given that it took about 30 years from the invention to the commercialization of lithium batteries, the acceleration potential through AI could prove decisive.
• More Computing Power: Finally, the construction of new and increasingly powerful data centers for AI calculations requires significant investment. According to the IEA, global investments in data centers have nearly doubled since 2022, reaching USD 500 billion in 2024. Ambitious, state-supported projects like "Stargate" in the USA provide additional momentum.

Why Invest in the Connection Between AI and Energy with Thematic Funds?

Selecting from this broad spectrum of investment opportunities may prove challenging for some investors. Thematic funds that align AI's potential with long-term trends can be a suitable instrument here. They are managed by expert teams and help to support both investment opportunities and risks broadly. These risks include not only the mentioned emission burdens but also technology and market risks.

The sustainable investment strategy "Swisscanto (LU) Equity Fund Sustainable Climate" focuses on the intersection of AI and energy, particularly in the industrial goods and IT sectors, as well as the utility industry. The focus is on applications that can lead to lower CO₂ emissions through the use of AI.

Three Key Points for Investing in AI and Energy:

Investment theme «Climate»: Insights