How Big Tech intends to power AI’s thirst for energy

The massive increase in the number of data centers around the world shows no signs of slowing down, prompting Big Tech companies to consider how best to drive the artificial intelligence revolution.

Among the options on the table are nuclear power, liquid cooling for data centers, and a pivot to quantum computing.

But critics say tech giants should recognize the cost of generating electricity as the pace of efficiency improvements in electricity use slows. A.I. Create a boom across the supply chain and let go of the “move fast and break things” narrative.

“Right now, the real environmental costs are completely hidden; they’re just subsidized by the fact that technology companies need products and buy-in,” said Head of Global Issues, Climate and Systems at Stockholm University. says Soumya Joshi, The Environmental Research Institute (SEI) told CNBC via video call.

The wave of data center investment expected According to the International Energy Agency, this trend will further accelerate in the coming years, primarily due to increased digitalization and the introduction of generative AI.

This outlook raises concerns about the rapidly increasing demand for electricity and the often overlooked but critically important impact of AI on the environment.

Data centers consume an ever-increasing amount of energy and are a critical part of the infrastructure supporting modern cloud computing and AI applications.

Giampiero Frisio, president of electrification at Swiss multinational ABB, said the engineering group’s data center business has experienced impressive growth in recent years, with the sector expected to grow by more than 24% in 2024. said.

Friscio said ABB is well-positioned in the AI ​​demand boom, supplying mid-sized and large industry players with all the components needed to run data centers.

“I think the best way to act now is to be more energy efficient. That’s the best way to go, because we have technologies like medium-voltage HiPerGuard UPS. We can do it, and we can do it tomorrow morning. You can do it,” Friscio said. via video call with CNBC.

HiPerGuard UPS refers to ABB’s industry-first medium-voltage uninterruptible power supply, which the company says can provide continuous power to large facilities.

“The second is moving to liquid cooling, no question about it. Again, this is for energy efficiency. Why? A single rack has all the servers. Because we have a black box that’s like a wardrobe.’The power density inside will be four to six times more than before,’ Friscio said.

“Then we’re talking about five to 10 years from now, and that’s nuclear modular systems,” he added.

US technology giant Microsoft, Google, and Amazon have all racked up billions of dollars worth of nuclear power in recent months as they seek to bring additional energy capacity online to train and run the large-scale generative AI models behind today’s applications. Secured an energy deal.

The surge in demand for generative AI coincides with a push to develop more efficient cooling solutions in data centers, particularly liquid cooling, a process that uses water to lower the temperature of servers and other electronic equipment.

French power equipment manufacturer Schneider Electric recently $850 million deal Acquires controlling interest in Motivair Corp., a US-based company specializing in liquid cooling for high-performance computing.

Schneider Electric’s CEO at the time told CNBC that the all-cash deal, aimed at increasing data center service offerings, was “expensive, but not overly expensive,” and that the company “It fits well” with the company’s strategy, he said.

Along with nuclear energy and liquid cooling technology, some tech companies have suggested that developments in AI could help decarbonize data centers.

For example, former Google CEO Eric Schmidt said: said Last month, he said investments in AI could be critical to solving some of the biggest environmental challenges because “we’re not going to meet our climate goals anyway.”

SEI’s Joshi flatly rejected this perspective.

“These arguments are not new and are very consistent with some kind of ‘silver bullet’, ‘technology will save us’ rhetoric,” Joshi said.

“There is an inherent contradiction in saying that we operate within a finite range.” planetary boundaries But beyond that, by continuing the same extractive narrative, we will somehow solve the problems we are currently in,” she added.

Raj Hazra, CEO of Quantinuum, the world’s largest integrated quantum computing company, said: “I think there’s a summer for every great technology we discover, but we don’t pay attention to it until winter arrives.” I’d rather not have it,” he told CNBC in a video call. .

“This is my way of explaining what’s happening with generative AI, the infrastructure needed to support it. [and] It’s a large data center that has to be built. ”

Hazra said optimism about the generative AI boom is already putting a strain on the technology’s operating costs.

“AI is great, but there are two challenges to it. One is, is it sustainable from a resource perspective? The other is, is it responsible?” Hazra said. Ta. “The reason I bring this context up is because quantum is important for both.”

Quantum computing refers to the field of computer science that uses the laws of quantum mechanics to solve highly complex problems.

Hazra said a “so-and-so” group of companies and strategic investors have recently expressed interest in Quantinum, and the company is raising a variety of funds. 300 million dollars In its latest equity funding round. Honeywell, the company’s major shareholder, said The funding brings Quantinuum’s valuation to $5 billion.

“One of the things that’s become pretty clear is that it’s no longer OK to say I have a solution to a problem. It’s no longer OK to say I have a sustainable solution to a problem. We have to,” Hazra said.

The CEO said one of Quantum’s biggest contributions to society is making AI sustainable and responsible.

“In the next three to five years, people will say, what is the computing infrastructure to run a business? It’s going to be a combination of high-performance computing, AI, and quantum. ” he added.