New data on China’s DeepSeek AI model suggests that more efficient models could lead to a surge in demand for data centers, contrary to initial concerns. UBS analysts now predict sector revenues will grow 20% in 2025, driven by improved efficiency gains.
Experts say cheaper and more efficient AI models could become an accelerator for the market, as they reduce power consumption and make it more accessible. Bruce Owen, EMEA president at Equinix, expects this trend to shape demand for data centers, which have traditionally been seen as a critical infrastructure for powering digital transformations.
However, some analysts caution that deeper analysis is needed to understand the impact of DeepSeek’s R1 model on market dynamics. Goldman Sachs’ research department forecasts a tighter balance between supply and demand in the coming years, with a peak expected in late 2026.
Efficiency gains may mitigate long-term market oversupply concerns, according to James Schneider, senior equity research analyst at Goldman Sachs. Meanwhile, investors are taking a wait-and-see approach on DeepSeek’s impact, with some stocks experiencing temporary downturns before recovering.
Experts agree that more efficient technology will ultimately lead to increased adoption of AI and drive up data center demand. Ryan Cox, head of AI at Synechron, expects efficiency gains to fuel further growth in the market, despite ongoing complexities and uncertainties.
Source: https://www.cnbc.com/2025/02/14/how-chinas-deepseek-could-boost-the-already-booming-data-center-market.html