tech
The AI industry is turning to an old energy favorite to bring its data centers online fast
AI companies are in a race to secure energy for their data centers. They are finding that natural gas is the easiest solution, if not the cleanest.
TL;DR
- Data centers are growing larger and more power-intensive, straining existing power infrastructure.
- AI companies need power quickly, with timelines of two to three years for data centers and four to eight years for grid upgrades.
- Natural gas offers a faster solution by utilizing existing pipeline networks and allowing for quicker plant construction or expansion compared to nuclear or renewables.
- Meta, Chevron, Engine No. 1, and Google are all investing in natural gas power for their data centers.
- Renewable energy sources, while long-term goals, are currently less attractive due to slower deployment and cost considerations, despite past cost reductions.
- The need for massive capital investment in AI infrastructure, with limited initial revenue, is driving the focus back to natural gas for economic reasons.
- Carbon capture technology is being considered as a way to mitigate the environmental impact of increased natural gas use.
- The current global 'structural supercycle,' driven by data centers, electrification, and cooling demand, is pressuring energy supplies, making natural gas a quick option.
- Natural gas produces less carbon dioxide than coal or oil but is still a fossil fuel contributing to the climate crisis.