Solar and wind PPA values in ERCOT fell about 4% this week amid uncertainty regarding data center energy demand. The downturn paused a 10% rally since September that was fueled by bullish load growth predictions.

On Monday, Chinese technology company DeepSeek introduced an AI model that performs nearly as well as models from OpenAI and Google but with less computing power and, in turn, lower energy needs.

Shares of companies powering data centers – like Constellation Energy and Vistra – fell by more than 20% on the news. The market’s reaction seemed clear: if the AI boom is no longer tied to energy-guzzling data centers, then the companies powering data centers might not be as well positioned as previously thought.

But some experts disagree that more efficient AI will lower future energy demand. After comparing the performance of DeepSeek’s technology to existing AI models, Scott Chamberlin of New Energy Efficiency predicted that, “these innovations are going to result in the same energy consumption growth trends.”

Key to this view is Jevon’s paradox, the notion that improved efficiency leads to increased consumption. In other words, if AI becomes more efficient, society will likely use AI for more things, and AI energy demand will continue to grow.

Other recent announcements support data center energy growth. Last week, President Trump announced Stargate, a $500 billion investment in private sector AI infrastructure –starting with data centers in Texas. Meta announced its own data center investment, too.  

Meanwhile, Goldman Sachs released a report finding that renewables could meet up to 80% of data center load by 2030.“Our conversations with renewable developers indicate that wind and solar could serve roughly 80% of a data center's power demand if paired with storage, but some sort of baseload generation is needed to meet the 24/7 demand,” Goldman’s Jim Schneider wrote.