AI leads to higher energy consumption and new infrastructure needs

TECHNOLOGYAI leads to higher energy consumption and new infrastructure needs

The development of artificial intelligence and its ever-increasing use are driving the generation and processing of data. Artificial intelligence solutions allow for the optimization of processes, improving competitiveness and reducing costs, while at the same time consuming a great deal of energy and generating significant amounts of carbon footprint. How does this happen? More data means larger IT resources are necessary for storing and processing AI data. The IT infrastructure designed for artificial intelligence consumes much more energy than traditional hardware solutions.

The Era of Artificial Intelligence

According to a McKinsey analysis from August of last year, the adoption of artificial intelligence has doubled in the past five years. The year 2023 was a breakthrough for this technology. In over half (56%) of the nearly 2,000 companies surveyed, artificial intelligence is used to support at least one function or task. Most respondents noticed a positive impact of AI on their organization, which translated into improved business efficiency or cost reduction. Over two-thirds of those surveyed expect their spending on AI-based tools to increase in the coming years.

Expectations for the development of artificial intelligence are also reflected in investments, particularly private ones. According to the Stanford “AI Index 2023” report, these have increased eighteen-fold in the last decade. And this is just the beginning, as record investments are expected in the coming years. Market research firm IDC forecasts that by 2025, large companies from the Forbes Global 2000, i.e., the world’s largest stock companies, will allocate 40% of their IT budgets to AI-related projects. The revolution will also affect the service sector, which is expected to use AI in 40% of its orders by 2025.

Artificial intelligence is supported not only by private investors but also by whole economies, national or community-based. A good example is the European Union, which plans to invest 1 billion euros annually in this technology. The implementation of long-term programs “Horizon Europe” and “Digital Europe” is expected to play a significant role here. According to the European Commission, such support can help mobilize the private sector and member countries, driving further investment, which is projected to be as high as 20 billion euros annually. Another EU incentive for European businesses to develop artificial intelligence is the AI Act, which is expected to come into force before this year’s European Parliament elections. The AI Act aims to guarantee citizens’ rights while encouraging innovation in artificial intelligence.

AI will change the data center market

One industry significantly influenced by the AI trend is the data center sector. Wojciech Stramski, CEO of Beyond.pl, a data center, cloud and Managed Services provider, believes the impact will be felt this year and in the coming years – “There are sweeping changes ahead in the area of data center design and energy consumption. Maintaining IT infrastructure for AI solutions requires an entirely new approach. Traditional solutions that we maintain at Beyond.pl data centers consume between 4 and 25 kW/h per rack cabinet. AI solutions have a demand of up to 100 kW/h per rack cabinet. This means that the majority of currently available server rooms are technically unable to handle such power. As a result, data centers accommodating the demand associated with the development of artificial intelligence will need to adopt new cooling solutions and change the equipment in their rooms. Beyond.pl, in response to such specific requirements, launched colocation services for AI, machine learning and high-performance computing solutions in new chambers of the Data Center 2 facility in January this year.”

Green Regulations

Considering the dynamically developing IT market, also driven by AI technologies, the European Union is effectively implementing policies to protect the environment. Member states will need to adjust their law to the guidelines recorded in the Energy Efficiency Directive adopted by the European Parliament by October 2025. It is already known that data centers will be obliged to report CO2 emissions, energy and water consumption, and consider the impact of their activities on the environment in planned investments. As part of the “Fit for 55” program, the EU is striving to reduce greenhouse gas emissions. More EU regulations are on the horizon, requiring organizations to be more energy-efficient. The obligation to report non-financial information related to the CSRD Directive enters into force for some entities this year. Organizations will need to monitor and report environmental indicators, including those related to IT infrastructure, to adapt to the new requirements.

“In the case of cooperating with a professional data center or cloud computing services provider, energy efficiency and water efficiency of the provider will be the key reporting factors.” explains Wojciech Stramski. – “In regards to data centers, this is indicated by the PUE (Power Usage Effectiveness) index, which is the ratio of total energy used by the facility to the amount of energy used to cool IT equipment, and the WUE (Water Usage Effectiveness) index, which shows the amount of water consumed in relation to the kilowatts of energy used to cool IT equipment. The average PUE value in Poland is 1.4-1.6, while the Data Center 2 facility belonging to Beyond.pl features a PUE of 1.2, one of the best energy efficiency indicators not only in Poland but also in the region.”

The type of energy purchased by the provider will also be essential. The use of renewable energy to maintain company IT infrastructure significantly reduces its carbon footprint.

Companies commercially developing artificial intelligence or using it for their own needs should, therefore, choose providers that ensure not only the proper infrastructure and specified computational power but also the proper cooling and scalability possibilities. It’s also very important to have guarantees for energy efficiency and water consumption, and the use of renewable energy, to reduce the carbon footprint. This is especially true given that the development of projects based on this technology is expected to be very dynamic.

Exit mobile version