AI’s Climate Promise Is Colliding With Its Energy Costs
AI is drawing climate-tech investment while increasing data center emissions, natural-gas demand, and public scrutiny over its environmental costs.
Artificial intelligence has become a central issue at New York Climate Week, where policymakers, investors, advocates, and journalists are gathering alongside the UN General Assembly. The debate is no longer limited to what AI can do. It now includes a more difficult question: Will the technology accelerate climate progress, or will its rapidly growing demand for energy make the crisis worse?
The answer remains unsettled. AI is attracting investment into energy technologies and could support scientific research, but the data centers behind it are also contributing to rising emissions, new natural-gas infrastructure, and resistance from affected communities.
Climate goals raise the stakes
The AI debate is unfolding as the world’s climate targets become harder to reach. A UN Environment Program report found that the world has nearly passed the point where warming could be kept below 1.5 °C above preindustrial levels.
That outlook increases the urgency of two parallel tasks:
- Rapidly reducing greenhouse-gas emissions
- Removing carbon dioxide that has already entered the atmosphere
AI could potentially assist with parts of this work, but it could also add to the emissions that climate policies are meant to cut. UN Secretary-General António Guterres summarized the tension during the opening day of the General Assembly: “Artificial intelligence could help solve all these challenges, or it could make them worse.”
That uncertainty is shaping conversations across the climate sector. Supporters emphasize AI’s potential to advance research and attract money to energy innovation. Critics point to the immediate environmental costs of building and powering the infrastructure it requires.
Data centers are reshaping climate investment
The expansion of AI is increasing demand for electricity, reliable grids, and new sources of power. That demand has brought greater attention and funding to energy companies, including some working on low- and zero-emissions technologies.
Startups in nuclear, geothermal, wind, and solar power have signed agreements with technology companies such as Google and Meta as those businesses look for energy to operate new or expanding data centers.
Venture-capital investment in climate technology reached $26 billion worldwide during the first half of 2026, according to industry finance tracker Currence. That was a 55% increase from the previous year. Products and services designed for data centers received a substantial portion of that funding.
The increase does not mean all climate technologies are benefiting equally. Investment in carbon management and low-carbon fuels fell sharply during the same period. These technologies may be important for addressing climate change, but they do not necessarily offer the direct, immediate value to data centers that power-generation and infrastructure companies can provide.
Big Tech’s emissions are moving in the wrong direction
The data center expansion has already created a significant emissions burden. Microsoft, Google, and Meta previously announced ambitious goals for reducing greenhouse-gas emissions, yet all three companies have since reported increases. The growth of data centers needed for AI has been a major factor.
Some of the new electricity demand may eventually be met with low- or zero-emissions power. In the near term, however, natural-gas infrastructure is being added to serve data centers. That creates a long-term concern because gas-fired power plants can remain in operation for decades after they are built.
The result is a mismatch between AI’s possible future climate benefits and the infrastructure being constructed to support it today. Even if cleaner energy ultimately powers more data centers, fossil-fuel projects approved during the current expansion could continue producing emissions well into the future.
AI may still support climate research
Despite the risks, there are reasons for cautious optimism. AI could help researchers move more quickly in areas such as the search for new catalysts, according to Evelyn Wang, MIT’s vice president for energy and climate.
Wang has also argued that data centers will not add to climate and water problems indefinitely. She estimated that it could take about a decade before data centers stop adding to planet-warming emissions.
AI’s climate value will therefore depend on more than improvements in software. It will also depend on how quickly electricity supplies become cleaner, whether grids can reliably handle rising demand, and whether funding reaches climate solutions that do not have an obvious data center customer.
The technology’s research potential does not automatically offset the emissions associated with its deployment. Any assessment must consider both the applications AI could enable and the physical systems required to run it.
Communities and regulators want clearer benefits
Public resistance is increasing in areas near data centers and the power plants that supply them. Residents are encountering concerns that include pollution and noise, while the broader climate sector remains skeptical about whether AI’s benefits justify its environmental footprint.
UN climate chief Simon Stiell warned that AI companies risk losing public support. He called on technology leaders to demonstrate that the benefits of AI outweigh its rapidly rising costs and that those benefits extend beyond a small group of companies or users.
This scrutiny puts pressure on the industry to provide evidence rather than broad promises. Companies will need to address local environmental effects, rising corporate emissions, and the consequences of new fossil-fuel infrastructure while making the case that AI can contribute meaningfully to climate solutions.
A climate test for the AI industry
AI is bringing new money and urgency to parts of the energy sector, but it is also redirecting investment and increasing demand for power. Its ultimate climate impact will depend on whether useful applications and cleaner infrastructure can develop fast enough to outweigh the emissions, pollution, and community disruption created during the current buildout.
For now, Climate Week’s dominant technology story is also an unresolved climate question: whether AI becomes a practical tool for decarbonization or another obstacle to achieving it.
Source: revew
Originally reported by revew.