SCANVAC
  • Home
  • About Scanvac
    • Board / Adm.
    • History
  • Blog
  • Scanvac Awards
  • Previous conferences
  • Co-operation
    • Nordic Ventilation Group (NVG) >
      • NVG Blog
      • Model IEQ regulation
      • Criteria of room air cleaners for particulate matter
      • Effects of indoor air humidity
      • Post-COVID target ventilation rates
      • Demand Controlled Ventilation
      • Control of airborne infections with ventilation and air distribution in post COVID pandemic
    • Ventilation regulations

Artificial intelligence accelerates data centers – energy infrastructure, waste heat and new processes in focus

13/10/2026

 
by Professor Risto Kosonen, Department of Civil Engineering, Aalto University, Espoo, Finland

​
The widespread adoption of artificial intelligence is reflected in the growth of data center projects. Despite weak market conditions, data center construction is booming. Public data indicate that roughly 200 major data center projects in the Nordic region are at some stage of planning. There is high demand for designers and, in particular, electrical contractors. Furthermore, rising electricity consumption requires additional generation capacity, which is primarily being pursued through renewable energy sources. Electricity transmission capacity is estimated to be a bottleneck. While wind power projects and data centers can be built relatively quickly, the implementation of new transmission lines takes years. Driven by growing demand for electricity, this could also raise the average price of electricity. At the same time, dynamic fluctuations in electricity prices will increase, creating a greater need for energy flexibility in buildings.

The proliferation of data centers offers an opportunity to electrify district heating and reduce its emissions, provided the centers are located near urban areas and their district heating networks. However, many projects are located in sparsely populated areas where heat demand is low, leaving waste heat unutilized. Even in urban areas, a challenge in utilizing data center waste heat is its low temperature level. Harnessing this heat requires heat pumps to raise the temperature to a level suitable for existing district heating systems. In new areas, district heating and building heat distribution systems can be designed for lower operating temperatures. In those cases, domestic hot water becoming the limiting factor, while a temperature of 60°C is required to mitigate the risk of Legionella. The return fluid temperature in the latest liquid-cooled data centers is already closer to this domestic hot water temperature, meaning heat pumps need to raise the temperature only slightly.

High-performance, next-generation data centers enable the large-scale use of artificial intelligence. AI is generally adopted with the aim of reducing costs and improving efficiency. Currently, however, AI is primarily used to enhance the efficiency of individual tasks. This is evidenced by a study of Nordic companies, which found that 74% of corporate executives view the value of AI chiefly in terms of improving efficiency in specific tasks.

In the long run, companies must focus on generating added value for their customers. In this context, new processes are needed to support growth and differentiation while creating genuine value. It is worth noting that isolated efficiency gains are often easily replicated, meaning the resulting advantage is short-lived. While efficiency measures do create value for the industry as a whole, competitors can easily copy the solutions, thereby diminishing a company's ability to differentiate itself. Ultimately, it is the paying customer who primarily reaps the benefits of these efficiency improvements. In contrast, genuine value arises when companies make their products or services superior to those of their competitors. An AI adoption strategy should be grounded in long-term competitiveness goals. This entails transforming existing processes to maximize the benefits derived from artificial intelligence.

It has been estimated that 10–15% of current jobs could disappear due to artificial intelligence. Technology replacing human labor is nothing new. During the Industrial Revolution, machine power replaced physical labor, and automation has since done the same. Now, for the first time, AI will significantly impact expert-type work, with effects varying greatly depending on the specific role. It has been proposed that job tasks can be broadly categorized into "bits" and "atoms." Roles involving "bits" are susceptible to being replaced by AI, whereas those working with "atoms" are more secure. Routine work involving "bits" generally generates little added value, while tasks involving "atoms" include, for instance, work at the customer interface and at the strategic level. This trend will make it difficult for young people to enter the field in the future, as they lack the necessary broad experience and insight.

However, it should be noted that practical installation work cannot be replaced by artificial intelligence. Plumbers and HVAC installers will still be needed even as the use of AI becomes widespread. At the same time, the standing of this profession is likely to rise.

In building services engineering, the use of artificial intelligence can be categorized into the application of language models and various data analytics methods. Language models can be utilized, for instance, to draft equipment lists and requests for proposals. These use cases often rely heavily on existing processes. Solutions have been developed to read human-readable documents—such as PDFs—and convert them into a format that computers can understand. This approach focuses on streamlining individual tasks. However, achieving a genuine boost in productivity requires transforming processes by starting with the semantic form of the data content. Semantic data content consists of information regarding meaning and relationships linked to a data model, enabling computer comprehension. The use of semantic models requires the standardization of data content—a necessary step for significantly improving process efficiency.

The use of data analytics methods has become increasingly common in areas such as system performance monitoring, fault diagnostics, and optimal control. These methods enable reduced energy consumption, the detection of anomalies, and cost savings. However, everything relies on measurement data collected from technical systems. Poor data quality has been observed in many contexts, and in such situations, the use of artificial intelligence fails to yield the desired results. Furthermore, a challenge for users is that many AI-based systems operate as closed "black box" models. This means that verifying their performance during actual use is difficult, and it is impossible to confirm whether the algorithm is functioning as intended.

Adopting new solutions requires radical innovations rather than incremental development. Otherwise, the benefits remain limited if one relies too heavily on old operating models.

“The people who are crazy enough to think they can change the world are the ones who do.” — Steve Jobs
<<Previous

    Blog of the SCANVAC President

    — a space dedicated to sharing reflections, perspectives, and updates from the heart of the Nordic HVAC community. Here, the President offers commentary on key developments in building services engineering, sustainability, indoor climate, and energy efficiency across Scandinavia and beyond.

    Archives

    October 2026
    March 2026
    January 2026
    December 2025
    November 2025
    October 2025
    September 2025
    January 2020
    March 2019
    June 2018
    December 2016

    RSS Feed

Proudly powered by Weebly
  • Home
  • About Scanvac
    • Board / Adm.
    • History
  • Blog
  • Scanvac Awards
  • Previous conferences
  • Co-operation
    • Nordic Ventilation Group (NVG) >
      • NVG Blog
      • Model IEQ regulation
      • Criteria of room air cleaners for particulate matter
      • Effects of indoor air humidity
      • Post-COVID target ventilation rates
      • Demand Controlled Ventilation
      • Control of airborne infections with ventilation and air distribution in post COVID pandemic
    • Ventilation regulations