Climate Action

How data centres and new power semiconductor technologies can support decarbonization

Data centres already heavily impact the world's carbon footprint, accounting for 1-1.5% of global electricity consumption in 2022.

Data centres already heavily impact the world's carbon footprint, accounting for 1-1.5% of global electricity consumption in 2022. Image: Reuters/Pavel Mikheyev

Adam White
Division President Power & Sensor Systems, Infineon Technologies
Our Impact
What's the World Economic Forum doing to accelerate action on Climate Action?
The Big Picture
Explore and monitor how SDG 07: Affordable and Clean Energy is affecting economies, industries and global issues
A hand holding a looking glass by a lake
Crowdsource Innovation
Get involved with our crowdsourced digital platform to deliver impact at scale
Stay up to date:

SDG 07: Affordable and Clean Energy

  • Every action we make online generates more data, meaning data centres play a critical role as the backbone of the internet.
  • Data centres already heavily impact the world's carbon footprint, accounting for 1-1.5% of global electricity consumption in 2022.
  • In the future, innovation driven by power semiconductor technologies could enable data centres to improve their energy balance.

Whether we're streaming a video, gaming, using generative artificial intelligence (AI) , or simply sending an email, we are constantly generating new data.

Since 2010, the amount of data generated annually has grown year-over-year, starting at 2 zettabytes. The latest estimates show that approximately 328.77 million terabytes of data are created every day, which equates to 120 zettabytes per year.

By 2025, the volume of data is expected to increase to more than 170 zettabytes ­– a more than 145-fold increase in 15 years only.

Why data centres are the backbone of the internet

With all this data traffic, data centres are playing a critical role as the backbone of the internet. They must process these immense streams of data around the clock.

The ever-increasing demand for data requires seamless connectivity, higher bandwidth, and wide-area coverage, which brings challenging requirements for modern data and communications infrastructures.

We also see that AI, and generative AI in particular, will accelerate this data growth and, in turn, increase demand for electricity. The public sector, business and civil society have demonstrated and understood that data centres already have an impact on the world’s carbon footprint and that this is set to become more pronounced.

Again, in numbers: in 2022, data centres alone accounted for 1-1.5% of global electricity consumption, according to the International Energy Agency.

Global data centre electricity consumption amounted to 460 TWh (terawatt hours; 1TWH = 109 kWh) in 2022; this is equivalent to the energy needs of 153 million households. By 2025, the electricity consumption of data centres is predicted to account for as much as 3.2% of total carbon emissions.

All these staggering figures automatically draw attention to the energy efficiency of data centres. They are one of the most energy-intensive building types, consuming 10 to 50 times more energy per floor space than a typical commercial office building.

The highest electricity consumption in data centres is largely driven by powerful servers and cooling systems. Servers get very hot during their continuous operation and therefore must be cooled.

This cooling requires further use of energy, around the clock. Some 50% of the energy consumed by data centres goes into cooling and backup power.

Only robust, reliable, secure and energy-efficient data centres can enable high-speed data transmission on the one hand and efficient operations that go hand in hand with net-zero ambitions on the other.

At the same time, they need to be designed for high scalability to allow AI developers to expand their computing and data storage resources.

How can we meet the demands of AI on data centres?

The role of data centres is critical in driving the advancement and widespread adoption of AI technologies across various industries. Although energy consumption has always been a concern with any new technology, the complexity of AI operations means it requires even more resources.

AI data centres are equipped with high-performance servers that can handle the massive computational workloads required for training deep learning models or running complex algorithms, for example.

Have you read?

They also offer vast storage capacities to store and manage these datasets securely and efficiently. To handle this, data centres need to be constantly upgraded to meet these demands.

The key question is: How can high data processing performance be guaranteed for the emerging and rapidly accelerating AI environment, given the limited physical space for power supplies in data centres?

How to guarantee high data processing performance

Solutions to this challenge must extend from the grid entering the data centre to the core, the AI processor. Innovations driven by power semiconductor technologies enable data centres to improve their energy balance.

They become more efficient and, in turn, reduce their operating costs. Advanced power semiconductors make power supplies more efficient and reduce cooling requirements.

Worldwide, energy savings of around 48 TWh could be achieved with various types of these advanced power semiconductors. This corresponds to more than 35 million tons of CO₂ emissions, according to Infineon analysis.

Looking beyond the right microelectronics, changes to the low-voltage power supply architecture in data centres also have the potential to positively influence its energy balance. This can lead not only to a reduction in power losses, but also to an increase in power density, enabling a further increase in computing power.

Promising new materials for greater efficiency

Technological innovation and improvement can also be achieved through new materials.

Gallium nitride (GaN) is one of the latest materials for semiconductor technologies that helps increase efficiency and power density. GaN semiconductors have high electrical conductivity and switching efficiency, thus helping to reduce energy consumption and improve energy efficiency of servers and data centres.


What's the World Economic Forum doing about the transition to clean energy?

At the same time, these types of power semiconductors are smaller and lighter compared with conventional silicon transistor-based implementations. This is also an important factor as it leads to less space required for the power supplies and fewer devices overall in the data centre.

These kinds of materials can be a game changer as they pave the way for higher power density, higher efficiency, and lower complexity – all of which ultimately lead to lower CO2 emissions.

To conclude, here’s another calculation, to illustrate the decarbonization potential of these new materials: If all US data centres were equipped with GaN semiconductors, energy savings would amount to 3.85 TWh. This is equal to a reduction of 3.15 million tons of CO2 emissions every year.

Don't miss any update on this topic

Create a free account and access your personalized content collection with our latest publications and analyses.

Sign up for free

License and Republishing

World Economic Forum articles may be republished in accordance with the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International Public License, and in accordance with our Terms of Use.

The views expressed in this article are those of the author alone and not the World Economic Forum.

Related topics:
Climate ActionEnergy Transition
World Economic Forum logo
Global Agenda

The Agenda Weekly

A weekly update of the most important issues driving the global agenda

Subscribe today

You can unsubscribe at any time using the link in our emails. For more details, review our privacy policy.

What’s so funny about climate change?

John Letzing

July 11, 2024

About Us



Partners & Members

  • Sign in
  • Join Us

Language Editions

Privacy Policy & Terms of Service

© 2024 World Economic Forum