AI is dominating the conversation at Climate Week
During the UN General Assembly and New York Climate Week, artificial intelligence became the dominant and most divisive topic among climate investors, policymakers, advocates and journalists gathered at panels and events. UN Secretary-General Antonio Guterres opened the assembly by framing the stakes directly: "The climate crisis fuels instability and displacement," he said. "Artificial intelligence could help solve all these challenges, or it could make them worse." His warning landed against a grim backdrop: a recent UN Environment Program report found that the world has nearly passed the point where it could still keep warming under 1.5 degrees Celsius above preindustrial levels, meaning the target is effectively slipping out of reach and both drastic emissions cuts and carbon removal are now needed.
AI's energy demand is reshaping climate-tech investment in the meantime. Global climate-tech venture capital investment hit $26 billion in the first half of 2026, according to data from industry tracker Currence, 55% higher than the same period last year, with products and services for data centers taking a large share. Startups in nuclear, geothermal, wind and solar have signed deals with companies including Google and Meta to power their growing data centers. But the boom is uneven: a Semafor piece on the same report noted that carbon management and low-carbon fuels, both important for cutting emissions but less attractive to data-center buyers, saw VC investment plummet this year.
The buildout has come at a real emissions cost. Microsoft, Google and Meta all set ambitious greenhouse-gas reduction goals a few years ago, but all three have since seen their emissions rise, largely because of the data centers needed to run AI. A wave of new natural-gas plants is also coming online to meet immediate power demand, and once built, those plants typically run for decades. Not everyone is pessimistic: Evelyn Wang, MIT's vice president of energy and climate, said during a panel that AI could help speed up progress in areas like the search for new catalysts, and told the Associated Press she expects data centers to stop adding to planet-warming emissions within about a decade.
Still, public patience appears to be thinning. UN climate chief Simon Stiell told the assembly this week: "AI leaders are now on thin ice when it comes to license to operate and sinking deep underwater when it comes to public support. Tech titans need to start showing why the benefits of AI outweigh its skyrocketing costs, for the many, not just the tiny few."
Key facts
- UN Secretary-General Antonio Guterres told the General Assembly that AI "could help solve" the climate crisis "or it could make them worse."
- A UN Environment Program report says the world has nearly passed the point where it could still keep warming under 1.5 degrees Celsius above preindustrial levels.
- Global climate-tech VC investment hit $26 billion in the first half of 2026, up 55% from a year earlier, per finance tracker Currence, with data-center-related products taking a large share.
- Microsoft, Google and Meta had set ambitious emissions-reduction goals a few years ago but have all since seen emissions rise, largely due to AI data centers.
- MIT's Evelyn Wang estimates it will take about a decade before data centers stop adding to planet-warming emissions; UN climate chief Simon Stiell said AI leaders are "sinking deep underwater when it comes to public support."
Why it matters
AI's energy appetite has collided head-on with a climate movement that just had to accept it has nearly missed the 1.5C target. The debate is no longer academic: UN leadership publicly cast AI as a technology that could go either way on climate, and framed the tech industry's credibility as already shaky with the public.
Who it affects
Climate investors and startups in nuclear, geothermal, wind and solar are benefiting from Big Tech's data-center power deals, while carbon management and low-carbon fuel startups, seen as less useful to data centers, are losing investment. Microsoft, Google and Meta are named directly as companies whose emissions have risen because of AI infrastructure, and communities near new data centers and gas plants are cited as sources of growing public pushback.
How to use it
This is a reporter's on-the-ground account of the debate at Climate Week rather than a product or policy announcement, so there is nothing to adopt directly. The practical takeaway for anyone tracking the sector is to watch two threads at once: climate-tech VC flows (up 55% year over year per Currence) and the emissions trajectories of the major cloud providers running AI infrastructure.
How solid is it
The piece is a first-hand newsletter dispatch from MIT Technology Review's climate desk, built on the author's own attendance at Climate Week panels and speeches, direct quotes from UN Secretary-General Antonio Guterres and UN climate chief Simon Stiell, investment data attributed to tracker Currence, a UN Environment Program report, and a cited Semafor analysis. No policy decision or new study is being announced, so it functions as an observational snapshot rather than breaking news.
Risks and caveats
The scale of the emissions increases at Microsoft, Google and Meta is not quantified in the source, nor is a mechanism given for how AI would speed up catalyst research. MIT's Evelyn Wang's decade-long timeline for data centers to stop adding net emissions is her own estimate, not a verified projection, and the natural-gas plants being built now are expected to operate for decades regardless of how AI's climate balance eventually shakes out.
“AI leaders are now on thin ice when it comes to license to operate and sinking deep underwater when it comes to public support. Tech titans need to start showing why the benefits of AI outweigh its skyrocketing costs, for the many, not just the tiny few.”
— Simon Stiell, UN climate chief