Yazılar

Japan’s METI Says DeepSeek’s Impact on Energy Demand is Hard to Predict

/in /tarafından

Japan’s Ministry of Economy, Trade, and Industry (METI) has stated that it is currently difficult to predict the potential impact of DeepSeek, a Chinese AI startup, on electricity demand. While there is a prevailing view that the growth of data centers could lead to increased power consumption, METI highlighted the complexities involved in determining how emerging technologies like DeepSeek will influence future energy requirements.

In December, the Japanese government released a draft of its updated basic energy plan, a policy document reviewed every three years. The plan projected a 10-20% rise in electricity generation by 2040, driven in part by the growing use of AI technologies. However, the advent of DeepSeek, which is rumored to consume less power compared to its competitors, has led analysts to debate whether electricity demand will rise or decrease. While some suggest that DeepSeek’s efficiency could lower demand, others believe that as the technology becomes more accessible and widespread, its adoption may ultimately increase power consumption.

METI explained that the relationship between AI and energy demand is influenced by numerous factors, including improvements in AI performance, cost reductions, and the development of energy-efficient technologies. As such, it remains challenging to predict how specific technologies like DeepSeek will affect Japan’s future energy needs.

The ministry emphasized that Japan’s economic growth and industrial competitiveness will be closely tied to these evolving dynamics, underscoring the importance of considering various technological, economic, and energy-related variables when forecasting demand.

 

Blackstone Remains Committed to Data Center Investments Despite DeepSeek Concerns

/in /tarafından

Blackstone reaffirmed its commitment to data center investments on Thursday, dismissing concerns that the rise of DeepSeek’s low-cost AI models would weaken demand for physical infrastructure. The alternative asset manager, which holds $80 billion in leased data centers, emphasized its “prudent approach” and strong partnerships with major global companies.

Data centers remain critical for AI development, providing the infrastructure needed to store, process, and analyze massive datasets. While investors previously saw data centers as key beneficiaries of AI growth, DeepSeek’s unexpected emergence has sparked debate over whether lower-cost AI models could reduce demand for such facilities.

Blackstone’s President and Chief Operating Officer Jonathan Gray addressed these concerns in a post-earnings call, stating that while the company is monitoring DeepSeek’s impact, lower AI costs could actually drive broader adoption, ultimately increasing data center demand. “As usage goes up significantly, there’s still a vital need for data centers. We still think it’s a very important segment,” Gray said.

Analysts at Jefferies echoed this sentiment, arguing that hyperscale cloud providers are unlikely to cut capital expenditures given the intensifying competition in AI. Tech giants such as Microsoft and Meta have also defended their aggressive AI spending, insisting that substantial investment is necessary to remain competitive.

Despite Blackstone’s confidence, its shares fell nearly 4% in afternoon trading, reflecting investor caution amid the evolving AI landscape.

 

ABB Confident in Data Center Growth Despite DeepSeek’s Energy-Efficient AI

/in /tarafından

ABB CEO Morten Wierod expressed confidence in the continued growth of the data center market, despite concerns over the impact of DeepSeek’s low-energy AI models. Speaking on Thursday after ABB reported its fourth-quarter results, Wierod reassured investors that demand for the company’s electrification products remains strong.

DeepSeek, a Chinese AI startup offering a more energy-efficient alternative to U.S. rivals, caused a selloff in tech stocks earlier in the week after surpassing OpenAI’s ChatGPT in downloads on Apple’s App Store. The AI model’s ability to operate with significantly fewer chips raised fears that it could reduce demand for data center infrastructure, impacting suppliers like ABB. ABB’s stock dropped nearly 6% on Monday amid these concerns.

However, Wierod said discussions with major partners and customers indicated that capital expenditure plans for data centers remain unchanged. ABB has significantly benefited from data center expansion, with its orders in this sector growing by an average of 23% per year between 2019 and 2023. The pace accelerated in 2024, with data center-related revenue now contributing 15% to ABB’s electrification business, up from 12% in 2023 and 8% in 2022.

While Wierod declined to provide a forecast for 2025, he emphasized ABB’s strong positioning in the industry, particularly in China. He also highlighted the company’s role in improving data center energy efficiency, an increasingly important factor as AI computing demand surges. ABB’s motors and variable speed drives can reduce electricity consumption by up to 60%, while its uninterrupted power supply systems operate at 97.4% efficiency when converting electricity.

The company also expects to benefit from the $500 billion AI infrastructure investment announced last week by U.S. President Donald Trump, reinforcing the long-term demand for data center solutions.

“The need for data centers and AI will be very strong in the coming years,” Wierod said. “I have no doubt.”