DeepSeek’s Chatbot Turns Out to be a Boon for Data Centers

Data centers are benefiting from China's DeepSeek artificial intelligence (AI) model's success, claims RHB Research.

While DeepSeek claimed to have trained its massive language model for under US$5.6 million, a fraction of the billions of dollars invested by US IT giants, the research firm noted in a report that this did not necessarily convert into a decreased demand for data centers. Rather, it stated that if DeepSeek's assertions were accurate, increases in AI efficiency may encourage wider use of AI, which would increase the need for additional data center equipment.

The research firm noted that this was consistent with the Jevons paradox, an economic theory that postulates that greater resource efficiency may result in higher total consumption.

RHB Research said that big hyperscalers including Microsoft Corp. and Meta Platforms Inc. have reaffirmed their commitment to significant expenditures in AI, even in the face of DeepSeek's rise. It further stated that Microsoft and Meta justified their US$80 billion and US$65 billion AI-related expenditures for the current fiscal year, arguing that it is essential to maintain long-term strategic competitiveness in the field.

The requirement for strong infrastructure to enable AI-driven services is anticipated to increase as the tech giants continue to invest billions in AI. According to RHB Research, Malaysia's cost advantages for labor, land, power, and closeness to Singapore makes it a desirable destination for data centers.