AI is being deployed far more to extract oil and gas than to help the climate, according to recent research, even as tech companies highlight selective clean energy projects at major climate conferences. A study published in a Nature journal in August found that the increase in carbon emissions from powering AI data centers and using AI to optimize fossil fuel production significantly outweighs the savings from using AI for clean energy applications.

Why Are Tech Companies Downplaying AI's Real Climate Impact?

During Climate Week NYC, major technology firms including Nvidia showcased how their AI products advance clean energy through projects like nuclear power plant optimization and battery storage systems. Yet researchers who study AI's environmental footprint say this narrative masks a much larger reality.

"What we're seeing at Climate Week is a familiar pattern: companies spotlighting a handful of clean energy pilot projects while the overwhelming majority of AI's actual applications are aimed at increasing fossil fuel extraction and production," said Holly Alpine, co-author of the Nature study and co-founder of the Enabled Emissions Campaign.

Holly Alpine, Co-founder, Enabled Emissions Campaign

Alpine noted that while renewable energy partnerships make for good publicity, companies rarely disclose how extensively AI is being deployed to find, extract, and produce more oil and gas. The International Energy Agency has confirmed that potential climate benefits from AI-enabled systems remain on the horizon, while concerning impacts like increased electricity demand for data centers are arriving much sooner.

What's Driving the Data Center Building Boom?

The rush to build AI data centers as quickly as possible has created a troubling infrastructure problem. Developers are relying on some of the least efficient and most expensive power technologies available, particularly natural gas combustion turbines that were originally designed for short-term, emergency use.

"It's madness," said Jonathan Koomey, an independent researcher who has studied data centers for decades. "No sane utility would buy electricity from combustion turbines running 24-7. It just wouldn't happen."

Jonathan Koomey, Independent Data Center Researcher

Koomey explained that projecting the scale of the data center boom even 10 to 12 years into the future is extremely difficult. One worst-case scenario involves utility companies building new power plants to serve anticipated data centers that never materialize, leaving households and businesses to cover the costs through higher utility bills.

How Can Organizations Actually Reduce AI's Energy Footprint?

While the challenges are significant, some companies are demonstrating that AI-powered energy management can deliver real results when applied thoughtfully. The key is shifting from reactive problem-solving to predictive optimization across buildings, data centers, and industrial systems.

• Smart Building Controls: AI-powered systems can analyze weather patterns and occupancy data to adjust heating and cooling in real time. One recent deployment across 616 retail stores saved over $1 million in costs within 12 months while reducing electricity consumption by almost 8 million kilowatt-hours.
• Data Center Cooling Optimization: Custom-engineered cooling solutions have reduced hyperscale data center energy use by 18 percent in India and cut over 16.5 million kilowatt-hours of electricity in Beijing, with corresponding reductions in carbon dioxide equivalent emissions.
• Predictive Performance Management: By combining real-time data with AI and efficient technologies, organizations can identify energy waste patterns that conventional systems miss, enabling incremental improvements over time rather than reactive fixes after problems occur.

The underlying principle across these applications is the same: understand how business needs and operational conditions impact energy requirements, then use that understanding to optimize demand proactively. This shift from reactive to predictive performance gives organizations new tools for managing energy consumption across residential heating, commercial buildings, and hyperscale data center operations.

What Happens When Economic Reality Meets AI Hype?

Experts warn that the current trajectory is unsustainable. Data centers will eventually need to reckon with the actual costs of their development and determine how much customers are willing to pay for their services. When that economic gravity asserts itself, the industry may face a significant downturn.

Regulators and lawmakers face mounting pressure to protect consumers from bearing the financial burden of speculative infrastructure investments. The challenge is that scaling energy-efficient AI solutions requires clearly defined goals, practical implementation pathways, and sustained human expertise to translate data insights into measurable performance improvements.

The gap between what tech companies promise about AI's climate potential and what the data actually shows remains stark. Until the majority of AI applications are powered by clean energy sources and emissions-cutting uses genuinely outweigh climate-harming uses, the selective storytelling at climate conferences will continue to obscure a more troubling reality.