The technology sector faces a major challenge: energy consumption and CO₂ emissions from AI and data centers are rising rapidly. Data centers already consume 2.1% of global electricity, and AI applications are pushing this demand even higher. Without countermeasures, the industry’s emissions could reach 14% of global emissions by 2040. This trend is echoed in recent analyses by the International Energy Agency (IEA), which warns that unchecked digital growth could have a significant climate impact.
Companies like Microsoft, Google, and startups such as Vilisto or Submer show how sustainable solutions can not only reduce the CO₂ footprint but also save costs. Now is the time to act and unite growth with climate protection.
The rising energy consumption of AI and data centers is becoming a real burden on the climate. The numbers speak for themselves and highlight how urgent it is to take action to get this growing energy demand under control. Recent studies, such as those from IEA and Nature, underscore the scale and urgency of the issue.
Data centers in the USA consumed about 200 terawatt-hours of electricity in 2024, with 53 to 76 terawatt-hours used by AI-specific servers alone. By 2028, electricity consumption for AI applications could rise to 165 to 326 terawatt-hours per year (IEA).
The global trend is equally alarming: Worldwide electricity consumption by AI data centers will rise from 50 billion kilowatt-hours in 2023 to around 550 billion kilowatt-hours by 2030—an elevenfold increase (Nature).
"As we move from text to video and images, AI models are getting bigger—and so is their energy demand. This will lead to significant energy consumption and further increase global emissions," explains Vijay Gadepally, senior scientist at MIT Lincoln Laboratory.
Data centers now consume 4.4% of total energy in the USA. In Europe, data center electricity consumption is expected to increase by 60% by 2030 (IEA).
Greenhouse gas emissions from data centers are also rising sharply: from 212 million tons in 2023 to 355 million tons by 2030. By 2030, global emissions from data centers could reach 40% of what the USA emits in a year (IEA).
Processing one million tokens with AI generates as much CO₂ as a gasoline-powered car driving 8 to 32 kilometers. Creating a single image with generative AI uses as much energy as a full smartphone charge (Nature).
"Global electricity demand from data centers will more than double in the next five years and by 2030 will be as high as Japan’s current electricity consumption," warns Fatih Birol, Executive Director of the International Energy Agency (IEA).
Additionally, reliance on fossil fuels further exacerbates the problem. According to Reuters, the carbon intensity of power used by data centers is often higher than the national average due to location and grid mix.
The carbon intensity of electricity used by data centers is 48% higher than the US average. This is because many data centers are located in regions with more carbon-intensive power sources or rely on quickly available fossil energy (Nature).
If the projected growth of US data centers through 2028 were entirely powered by natural gas, it could result in about 180 million tons of additional CO₂ emissions annually (IEA).
The impact varies by location: A 100-megawatt data center in Northern Virginia would emit about 463,000 tons of CO₂ per year, while the same facility in Texas would produce only 386,000 tons—a 17% difference (Utility Dive).
"AI data centers require a constant energy supply, 24 hours a day, 365 days a year," emphasizes Rahul Mewawalla, CEO of Mawson Infrastructure Group.
Currently, data centers consume 1–2% of global electricity. However, this share could rise to 3–4% by the end of the decade (IEA). This development makes it clear how urgently a switch to more sustainable energy sources is needed to reduce the tech industry’s environmental footprint.
Tech companies have numerous opportunities to significantly reduce their CO₂ footprint—through smart strategies and the use of modern technologies. The IEA and US EPA both highlight the importance of a holistic approach.
One of the most effective measures for sustainable data centers is energy-efficient cooling. Conventional air conditioning often uses more energy than the servers themselves. Technologies such as direct-to-chip cooling or immersion cooling offer solutions here, as they can significantly lower energy demand. The MIT Technology Review details how immersion cooling and other innovations are transforming data center efficiency.
For example: Google reduced its cooling costs by 40% using its DeepMind AI system. The AI analyzes factors such as temperature patterns, airflow, and cooling efficiency (DeepMind).
Microsoft also relies on smart technology: Predictive algorithms help dynamically schedule workloads, allowing servers to automatically switch to energy-saving modes. Meta follows a similar approach, using AI models to analyze emissions, heat, and airflow for maximum energy efficiency (Meta Sustainability).
Huawei improved the Power Usage Effectiveness (PUE) in Chinese data centers by 8% through predictive analytics. Additionally, optimizing server utilization through consolidation and workload rightsizing increases efficiency (Huawei News).
Beyond innovative cooling methods, switching to sustainable energy sources plays a central role.
Transitioning to energy sources like wind, solar, and geothermal is a crucial step in reducing emissions. Germany has taken a leading role: In the first half of 2024, 65% of public net electricity generation came from renewables (Clean Energy Wire).
Wind power contributed 73.4 TWh, accounting for 34.1% of net electricity generation. Photovoltaic systems delivered 32.4 TWh—a 15% increase over the previous year. Data centers can also strategically relocate workloads to sites where renewable energy is more available (IEA Renewables).
In 2023, Germany invested $89.2 billion in clean technologies, accounting for almost 29% of the EU total (BloombergNEF). These investments lay the foundation for sustainable digitalization. Starting in 2027, Germany’s Energy Efficiency Act will require data centers to be powered exclusively by renewable energy.
"Germany’s strong commitment to climate protection, renewable energy, and policy frameworks has set a benchmark that encourages industry leaders and other countries to accelerate their climate goals," explains Miranda Gardiner of the Infrastructure Connectivity Alliance.
In addition to Green IT and renewables, Lifecycle Assessments (LCAs) enable a comprehensive evaluation of the environmental impact of AI systems and data centers. LCAs consider the entire lifecycle, from hardware manufacturing to disposal. The UN Environment Programme also promotes LCAs as a tool for sustainable decision-making.
These analyses help companies identify critical environmental impacts and strategically improve processes. BASF, for example, uses LCAs to select and market products with the best environmental profiles (BASF LCA).
Germany’s Green IT initiative demonstrates how effective a structured approach can be: Between 2008 and 2021, the federal administration reduced its energy consumption from 649.65 GWh to 334.54 GWh—a 49% decrease, saving €546 million (German Federal Government CIO).
Moreover, LCAs support ESG reporting by providing data-driven insights for better decisions, sustainable development, and regulatory compliance (GRI Standards).
By combining energy-efficient hardware, AI-powered algorithms, and renewable energy, the environmental impact of AI operations can be significantly reduced. A holistic approach that considers all aspects of data center operations is essential.
Real-world examples show how companies are already successfully reducing their CO₂ footprint. These case studies illustrate how strategies like Green IT, renewable energy, and lifecycle assessments are implemented in practice—with measurable results.
Microsoft has committed to using 100% renewable energy for data centers and buildings by 2025. Between 2021 and 2022, they were able to reduce their Scope 2 emissions by 30–35%. See more about reducing Scope 2 emissions and in Microsoft’s sustainability blog.
Apple has used “Clean Energy Charging” since 2023, a technology that times charging to periods with low-emission electricity. This has noticeably reduced Scope 3 emissions in the USA (Apple Newsroom).
A leading tech company relocated parts of its production from Asia to North America in 2023. Shorter transport routes led not only to fewer emissions but also to faster delivery times (McKinsey).
Interestingly, 90% of tech companies have already defined decarbonization targets. Many report energy cost savings of up to 20%. Tesco reduced emissions per area by 41% and saved £37 million annually. Nike cut energy intensity in manufacturing by 50% (CDP).
There are also exciting developments internationally in sustainable data centers.
Submer from Spain uses immersion cooling, where servers are submerged in non-conductive liquids, reducing both energy consumption and water usage. The company raised $55 million in Series C funding (Data Center Dynamics).
ZutaCore from the USA offers waterless direct-on-chip cooling (HyperCool®), cutting cooling energy use by up to 50% while increasing computing power per area tenfold (ZutaCore).
CoolestDC from Singapore develops liquid cooling systems with a PUE below 1.06, enabling energy savings of 25–50% (CoolestDC).
Heata from the UK uses server heat to provide free hot water for households, saving up to £340 and avoiding 750 kg of CO₂e emissions per home annually (Heata).
More and more companies are using data-driven climate risk assessments to make strategic decisions. The 40 DAX companies reduced their greenhouse gas emissions by 4% in 2023 (14% excluding Scope 3). Learn about carbon stress tests for business resilience (CDP).
Companies excelling in growth, profitability, and ESG criteria—so-called “triple outperformers”—achieved annual shareholder returns 2 percentage points higher than purely financially focused firms (McKinsey).
Global ESG assets surpassed $30 trillion in 2022 and could reach $40 trillion by 2030 (Bloomberg). An example of confidence in this sector is the IPO of Haffner Energy, which raised €74.4 million (Reuters).
These figures and examples show: Sustainability is not only an ecological imperative but also an economic opportunity. By leveraging modern technologies, strategic planning, and consistent implementation, companies can protect the environment and benefit in the long term.
Tech companies face the urgent task of reducing their CO₂ emissions—for both environmental and economic reasons. Taking a close look at the entire lifecycle of digital infrastructure can help identify clear action points. The US EPA and IEA both recommend comprehensive lifecycle strategies.
Lifecycle Assessments (LCAs) capture the full environmental footprint of a product or service. Unlike traditional methods, LCAs cover all phases—from design and operation to disposal. This comprehensive approach reveals not only the true environmental costs of digital infrastructure but also often-overlooked opportunities for improvement (US EPA).
For a complete picture, LCAs must consistently include all phases. Data center operators should deploy cross-functional teams to meet new regulatory requirements. This requires new processes, targeted training, and technologies such as sensors and data center infrastructure management (DCIM) software solutions (Gartner).
"The process is slow to get started. You have to take it seriously and figure out how it works, and the industry has to take it seriously and realize that this data needs to be reported and figure out how to do it." – Jay Dietrich, Research Director for Sustainability at the Uptime Institute
The costs of renewable energy have fallen sharply in recent years: Between 2010 and 2020, solar prices dropped by 85%, while wind energy became 56% cheaper (IRENA). By 2030, renewables could cover up to 65% of global electricity demand (IEA World Energy Outlook).
Renewable energy sources emit only about 50g CO₂ or less per kilowatt-hour—a stark contrast to coal (1,000g CO₂/kWh) or natural gas (475g CO₂/kWh) (IEA Data). In addition, hybrid cooling approaches and heat recovery can boost efficiency and further reduce the environmental footprint (Data Center Dynamics).
Another benefit: Investing in renewables protects against fossil fuel price volatility and improves air quality. Around 80% of the world’s population lives in countries dependent on fossil fuel imports (IEA World Energy Outlook). The UK is a prime example: The share of carbon-free electricity sources there rose from under 20% in 2010 to 50.3% in 2023 (Carbon Brief).
"It’s time to stop burning our planet and instead invest in the abundant renewable energy all around us." – António Guterres, Secretary-General of the United Nations.
Successful ESG strategies (environmental, social, and governance) set clear targets, involve stakeholders, and continuously monitor progress. ESG principles should not be viewed as isolated measures but fully integrated into corporate strategy (GRI Standards).
A good starting point is a materiality assessment, which helps identify and prioritize the most important ESG factors. These considerations should flow into all decision-making processes, supported by regular monitoring and transparent reporting. Companies can look to best practices to strengthen their own strategies (McKinsey).
Such a systematic approach not only helps reduce the tech sector’s environmental footprint but also offers long-term competitive advantages.
The tech industry stands at a pivotal point. According to the International Energy Agency (IEA), global electricity demand from AI could double by 2026. At the same time, companies are already proving that sustainable growth is possible. Forecasts show that data center energy consumption could rise from 460 TWh in 2022 to over 1,000 TWh by 2026 (IEA). Yet this challenge also opens up opportunities for new solutions.
A prime example is the CloudHQ data center in Schleswig-Holstein. Thanks to direct connections to offshore wind farms, it can cut electricity costs by 30%. Equally impressive is BMW’s €70 million investment in AI infrastructure in Dingolfing: Energy consumption was reduced by 18% while throughput increased by 12% (BMW Group).
"The combination of surplus wind energy and limited transmission capacity offers a unique opportunity for energy-intensive AI workloads in northern Germany." – Hossein Fateh, CEO of CloudHQ
Legal frameworks are also driving change. Germany’s Energy Efficiency Act will require data centers to use only renewable energy starting in 2027. Major companies like Amazon, Microsoft, Meta, and Google are also active: Through corporate renewable energy power purchase agreements, they have secured nearly 50 GW of capacity—equivalent to Sweden’s entire electricity generation (BloombergNEF).
In addition to regulations, technological innovation plays a key role. AI can make power grids more efficient and optimize electricity distribution. Studies show such measures could cut global emissions by up to 4% by 2030 (IEA Digitalisation and Energy). At the same time, ambitious climate action could bring the global economy benefits of at least $26 trillion by 2030 (New Climate Economy).
Companies must develop future-proof strategies based on energy-efficient hardware, alternative cooling methods, and decentralized power generation. The results speak for themselves: knowmad mood increased its revenue by 19.93% and normalized EBITDA by 28.40% in Q1 2025 through sustainable measures.
The tech sector has the tools to combine climate protection with innovation. With renewables already accounting for almost 30% of global electricity production (IEA Renewables 2023) and a green bond market worth $270 billion in 2020 (Climate Bonds Initiative), the financial foundation is in place. The approaches presented here make it clear: Growth and climate responsibility can go hand in hand—if companies act now.
How companies in Germany can reduce the CO₂ footprint of AI and data centers
Companies in Germany have numerous ways to lower the CO₂ footprint of their AI applications and data centers. A key step is the use of renewable energy such as wind and solar power. These energy sources not only make operations more environmentally friendly but also help cut costs in the long run (Clean Energy Wire).
Additionally, optimizing server infrastructure plays a crucial role. Efficient cooling technologies and modernized systems can significantly reduce energy demand. The goal: more performance with less consumption. For example, MIT Technology Review describes how immersion cooling and AI-driven optimization can cut energy use.
Another approach is to use Green IT strategies, where AI itself is used to optimize energy consumption and operations. Companies building sustainable data centers and making smart use of waste heat make a vital contribution to achieving Germany’s climate targets by 2045.
These steps are not only good for the environment but also offer a competitive edge in a market increasingly focused on sustainability.
The use of renewable energy in data centers
Renewable energy is a key lever for shrinking the CO₂ footprint of data centers in Germany. Starting in 2024, at least half of the electricity used by data centers must come from renewable sources such as wind, solar, or hydropower. By 2027, this share is set to rise to 100% (Clean Energy Wire).
Besides reducing emissions, renewables also offer economic benefits. They help companies stabilize energy costs over the long term, as they are less susceptible to fossil fuel price fluctuations. At the same time, many companies are investing in Green IT and energy-efficient technologies to drive sustainable digitalization. Such investments enable companies to meet growing demand for computing power with a smaller environmental impact (IEA Renewables).
Technologies for reducing energy consumption in data centers
Data centers use various technologies and strategies to cut energy consumption. Real-time monitoring combined with sensors plays a key role in accurately capturing and analyzing energy demand. In addition, AI-powered systems optimize cooling operations, making them more efficient (MIT Technology Review).
Physical measures such as liquid cooling and hot/cold aisle containment also help dissipate heat more effectively. Complementing this, virtualization technologies and well-planned airflow management ensure better use of available resources (Data Center Dynamics).
These approaches not only help companies reduce their CO₂ emissions but also enable significant operational cost savings—a win for both the environment and business.