What happened
data center electricity demand natural is at the center of this update. Natural gas power plant construction costs have surged by 66% over the past two years, driven in part by skyrocketing electricity demand from rapidly expanding AI data centers. Build times have also increased by 23%, reflecting supply chain constraints and labor shortages. These plants are critical to powering the compute-heavy workloads that companies like OpenAI, Anthropic, and Elon Musk’s xAI require.
Why it matters
The AI industry’s growth hinges on reliable, cost-effective energy. The significant increase in natural gas power plant costs and delays threatens to constrain data center expansion, potentially slowing AI research and deployment. As these companies compete to lead the AI race, infrastructure bottlenecks could become a strategic disadvantage.
Context
AI models such as ChatGPT, Claude, and Grok demand immense computational resources, driving unprecedented electricity consumption in data centers. While renewable energy adoption is growing, natural gas remains a key power source in many regions, making its cost and availability critical to AI infrastructure development.
Expected impact
Rising power plant costs and longer construction timelines may increase operational expenses for AI companies and delay data center growth. This could push AI firms to explore alternative energy strategies or accelerate efficiency improvements to mitigate risks tied to natural gas dependency.
What we still do not know
The AI industry’s adaptation strategies to these energy challenges remain unclear. The balance between natural gas and renewable energy in future data center power mixes is uncertain, as is the potential impact on competitive dynamics among AI leaders.
Related coverage: AI Chronicle analysis and updates.

Andrej Karpathy Calls for an End to Policing AI-Generated Homework in Schools
Marble Secures $9 Million to Revolutionize Tax Work with AI Amid Accounting Labor Shortage
Top 5 Clawdbot Security Risks and Solutions to Watch in 2026
Meta Secures 1 Gigawatt of Solar Power to Support AI Data Centers and Reduce Carbon Emissions