More than the global tech market was shaken this week with the emergence of DeepSeek, a Chinese startup that not only flipped assumptions of AI development but also sent shockwaves through the energy and financial sectors.
The company’s breakthrough in AI efficiency has forced a reassessment of the idea that power demand would continue to expand as AI technology progressed.
For over a year, energy companies have been riding a wave of optimism, fueled by the belief that the AI revolution would drive unprecedented power demand. The logic was straightforward: training and running AI models requires vast computational resources, and data centers—already consuming over 1.5% of global energy and 4% of U.S. electricity—were projected to double their share in just a few years.
Energy firms aggressively expanded their investments, expecting a surge in AI-driven power consumption.
Then came DeepSeek’s R1, an open-source AI model that seemingly does more with less. Developed in just two months and trained at a fraction of the cost of its Western counterparts, R1 challenges the assumption that AI growth must come with an exponential increase in energy consumption.
The result was a huge shakeup in the price of tech stocks on Jan. 27, with Nvidia’s stock plummeting nearly 17% in a single day, erasing over $600 billion in market capitalization. Investors had long assumed that AI’s growth would sustain Nvidia’s dominance, with hyperscalers like Microsoft, Google, and Meta spending billions on its high-end chips.
But another casualty was energy stocks, with a historic selloff in U.S. energy stocks on Jan. 27 taking billions in market value from companies like Constellation Energy, Vistra, and Talen Energy losing billions in market value overnight.
The development has thrown energy market projections into disarray. If DeepSeek’s efficiency gains can be replicated, the anticipated power build-out for AI data centers may be significantly smaller than previously estimated.
“We still believe data centers, reshoring, and the electrification theme will remain a tailwind, but that market expectations went too far. We reaffirm our fair value estimates and High Uncertainty Ratings for Vistra and NRG,” Morningstar said on Jan. 27.
This story was written by Viktoriya Zakrevskaya.
