Under the AI revolution, in order to meet the growing energy demand, tech giants are competing to come up with solutions, including liquid cooling for data centers, nuclear energy, and quantum computing.
According to the International Energy Agency (IEA), the wave of investment in data centers is expected to accelerate in the coming years, primarily driven by digital growth and the proliferation of generative AI. This outlook has also raised concerns about the surge in electricity demand and the environmental impact of AI.
Some analysts believe that as the pace of improving electricity usage efficiency slows, tech giants need to recognize the costs associated with the generative AI boom across the entire supply chain and abandon the approach of &#34;acting fast and breaking the norm.&#34; Recently, Somya Joshi, head of the Global Agenda, Climate, and Systems Department at the Stockholm Environment Institute (SEI), stated in a CNBC video interview:
<blockquote>
&#34;Currently, the actual environmental costs are quite hidden. Because tech companies bear part of the environmental costs in order to launch products and gain market recognition.&#34;
</blockquote>
<h2>Continuous Power Supply, Liquid Cooling... Fully Enhancing Energy Efficiency</h2>
As the infrastructure behind modern cloud computing and AI applications, the energy consumption of data centers is increasing. Giampiero Frisio, President of Electrification at Swiss multinational ABB, stated that the company's data center business has seen significant growth in recent years, with an expected increase of over 24% in 2024.
He mentioned that ABB is well-positioned in the AI demand boom, able to provide all the components needed to operate data centers for medium and large industry players:
<blockquote>
“I think the best course of action now is to improve energy efficiency. Because the technology already exists, such as the medium-voltage HiPerGuard UPS (ABB's pioneering medium-voltage uninterruptible power supply), which can provide continuous power for large facilities.”
</blockquote>
Frisio also noted, “The second thing is to shift to liquid cooling, which is undoubtedly for better energy efficiency. Because a single rack, you know those black boxes that look like cabinets filled with servers, will have a power density four to six times that of previous models. Five to ten years from now, that will be nuclear modular systems.”
<h2>Tech Companies Entering Nuclear Energy to Support AI Development</h2>
In recent months, American tech giants Microsoft, Google, and Amazon have signed nuclear energy agreements worth billions of dollars, seeking to increase additional energy capacity to train and run the large generative AI models behind today's applications.
Previously, Google signed a power purchase agreement with Kairos Power for a small modular reactor (SMR) that will generate 500MW of energy by 2035. Amazon has signed agreements with utility companies in Virginia and Washington to support the development of SMR projects and plans to invest in SMR technology developer X-energy. Additionally, in July of this year, the U.S. Department of Energy announced $900 million in funding for SMR development
As the demand for generative AI surges, people are also looking for more efficient cooling solutions for data centers, particularly liquid cooling (the process of using water to lower the temperature of servers and other electronic devices). Recently, French electrical equipment manufacturer Schneider Electric completed an $850 million deal to acquire a controlling stake in the American company Motivair Corp, which specializes in high-performance computing liquid cooling.
The then CEO of Schneider Electric stated that this all-cash transaction aims to strengthen its supply to data centers, describing the deal as “huge in amount” but “not too expensive,” and “very fitting” for the company's strategy.
<h2>Quantum Computing Can Make AI Sustainable and Responsible</h2>
Quantum computing refers to a field of computer science that uses the laws of quantum mechanics to solve extremely complex problems. Raj Hazra, CEO of the world's largest integrated quantum computing company Quantinuum, stated that the optimism surrounding the generative AI boom has put pressure on the costs of running the technology:
<blockquote>
“I think every great technology has a summer and a winter, but don’t wait for winter to pay attention to it. This is how I describe what is happening with generative AI, and the infrastructure needed to support it, along with the massive data centers that must be built.”
</blockquote>
He believes that AI today faces two challenges: first, from a resource perspective, is it sustainable? The second is, is it responsible? He stated that quantum is crucial for both:
<blockquote>
“One of the greatest contributions of quantum to society may be making AI both sustainable and responsible. I predict that in the next three to five years, you will hear people asking, what is the computing infrastructure that runs my business? It will be a combination of high-performance computing, AI, and quantum.”
</blockquote>

003816

01816

601985

American tech giants Microsoft, Google, and Amazon have signed multi-billion dollar nuclear energy agreements to meet the energy demands of artificial intelligence. Analysis indicates that quantum computing can help improve the energy efficiency of AI. The International Energy Agency predicts that investment in data centers will accelerate, driving a surge in electricity demand. Experts urge tech companies to recognize the environmental costs of generative AI and improve energy efficiency by adopting technologies such as liquid cooling. ABB expects its data center business to grow by more than 24% in 2024

- The article discusses the increasing energy demands driven by the rise of AI, highlighting solutions like liquid cooling, nuclear energy, and quantum computing.  
- Major tech companies are investing heavily in nuclear energy agreements to support the energy needs of generative AI, with significant investments and partnerships announced recently.  
- Experts emphasize the importance of improving energy efficiency in data centers and the potential of quantum computing to make AI development more sustainable and responsible.

From nuclear energy to quantum computing, how do tech giants plan to meet the energy demands of artificial intelligence?