New York Climate Week — roughly 100,000 activists, diplomats, funders and journalists descending on Manhattan alongside the UN General Assembly every September — has historically been about emissions targets and timetables. This year, per CNN analysis published Monday (September 28), it was about something else entirely: how to power the AI revolution. “Most notably, the word ‘climate’ was curiously absent from many panel discussions,” wrote CNN’s Andrew Freedman, a decade-long Climate Week attendee, who observed “a shared recognition bordering on a resigned acceptance that we are missing the climate targets.” CNN’s read of the week: the economic need to scale data centres for AI “has stripped the climate challenge of its sense of urgency.”
🔍 THE BOTTOM LINE: The world’s biggest annual climate gathering became a data-centre power conference, and the people who built it noticed. The tension is not abstract — it is playing out in blocked projects, pivoted startups, and a backlash that is starting to cost real money.
The numbers behind the takeover
The AI pull on the agenda is measurable. Climate-tech venture investment hit $26 billion in the first half of 2026 — up 55% on last year — with MIT Technology Review reporting that “products and services for data centers are getting a massive slice of that pie.” Nuclear, geothermal, wind and solar startups are signing supply deals with Google, Meta and peers hunting clean electrons for their facilities.
The same reporting documents what is being crowded out: carbon management and low-carbon fuels saw venture investment plummet this year — solutions that matter for climate but, as MIT TR puts it, “may not be able to sell themselves to a data center.” UN Secretary-General António Guterres named the tension in his assembly address: “Artificial intelligence could help solve all these challenges, or it could make them worse.”
The cost of saying that out loud appears to be rising. TechCrunch’s Climate Week dispatch describes corporates who are “still interested in climate” but “don’t want to crow about it, mostly for fear of drawing the Trump administration’s ire,” and founders being bluntly asked where the money was three years ago. Two energy founders, asked whether they’d prefer the AI buildout at its current pace or a climate-responsible speed, answered without hesitation: faster was better, per TechCrunch. Both startups sold energy.
The backlash is already costing billions
The political resistance executives keep being asked about has a price tag. Bloomberg reporting from the same meetings cites research group Data Center Watch: roughly 45 developments worth $68 billion were blocked or delayed by local pushback in the second quarter of this year, and thirty US statehouses introduced measures on data-centre siting, electricity and water.
Google’s global head of data-centre energy, Amanda Peterson Corio, put the political position bluntly at a Bloomberg Green event: data centres are politically hated “on the left and the right” per Bloomberg, with elections coming and “a lot more money” about to be spent on the fight. Microsoft’s chief sustainability officer Melanie Nakagawa answered with the defensive toolkit — ring-fencing data-centre costs away from consumer bills, workforce investment, published water-usage data. Amazon’s Kara Hurst offered the statistic-of-the-week: 90% of the time Amazon is not using water to cool its data centres.
The emissions ledger underneath is not subtle. Microsoft, Google and Meta all had ambitious reduction goals a few years ago; MIT TR notes all have seen emissions rise, largely because of the data centres powering AI. And the immediate gap is being filled with natural gas — plants with decades-long lifetimes — which is precisely why the site has called AI’s energy problem the boom’s unresolved bill.
Why this year felt like a pivot point
Climate Weeks are pulse readings, and this year’s pulse said the centre of gravity moved. The funding shift is self-reinforcing: startups that can pitch grid infrastructure or dispatchable power raise $14 billion quarters; startups researching carbon removal or low-carbon fuels, which cannot “sell themselves to a data centre,” watch investors walk. Money is not just following AI — mission statements are. A founder community that spent a decade on carbon-cutting is now, in TechCrunch’s framing, treating the data-centre boom as a bridge that might “last long enough to help startups build durable businesses” before they refocus on the mission they were founded to pursue.
That is either pragmatic or corrosive depending on where you stand, and Climate Week hosted both readings without resolving them. What neither side disputes is the mechanism: AI’s electricity demand is now large enough to redirect capital, silence language and reshape a movement’s biggest annual event — a speed of influence that surprised even the people tracking it.
The New Zealand angle
New Zealand sits at one remove from this fight — no hyperscale data-centre politics yet, and an electricity grid already priced by dry-year risk rather than AI demand. But the pattern is arriving here in slower motion: Invercargill’s proposed 280MW data-centre made energy security the centre of its pitch precisely because that is what AI projects now sell, and Australian moratorium politics show where the backlash goes when communities feel the costs land before the benefits. The Climate Week lesson for NZ policymakers is procedural, not rhetorical: the jurisdictions that got ahead of siting, water and electricity rules are the ones where projects did not become $68 billion worth of wreckage. Ours are still quiet — which is the window in which rules are cheap.
What we still don’t know
Climate Week is a conversation proxy, not an outcome measure; no one has quantified how much actual climate-policy attention was displaced. The Data Center Watch figures are quarterly and US-only, and the group’s methodology has not been independently audited. Whether the pivot-to-AI funding pattern survives the buildout’s current pace — or leaves climate-tech’s unfashionable sectors permanently underfunded — will only show in next year’s numbers. And the UN’s own read, that AI could help or worsen the climate crisis, remains exactly the unanswerable question everyone in New York this week was, in one way or another, being paid to have an opinion about.
📰 Sources
- CNN — “How AI took over a key climate event” (September 28, 2026)
- TechCrunch — “The AI boom took over Climate Week and not everyone is happy about it” (September 28, 2026)
- MIT Technology Review — “AI is dominating the conversation at Climate Week” (September 24, 2026)
- Bloomberg via Energy Connects — “Microsoft, Google Confront Rising Data Center Backlash” (September 23, 2026)
