New York, frenetic at the best of times, bordered on frantic when Climate Week coincided with the U.N. General Assembly meeting in September.

While the U.N. addressed climate in its hushed halls, experts and pundits at hundreds of Climate Week events scattered across dozens of locations analyzed every aspect of it. Like a subway running beneath this melee of meetings, there were common themes that connected the many panels and events I attended: Old goals were quietly dropped, new challenges accepted and an AI-led future imagined.

Note: Many panels were under the Chatham House Rule or on background only, so some ideas and quotes below are not attributed. Just know that all the speakers had lengthy titles, impressive biographies and well-known employers.

Hello ChatGPT, Goodbye 1.5 Degrees

This was my first climate event where there was little mention of the Paris Accord goal of limiting the global average temperature increase to 1.5 degrees Celsius above pre-industrial levels. “One-point-five” used to be a term so common it was rarely explained. IYKYK — and Climate Week attendees know. This year, the few times it was used, it was in a casual aside about the target there was no chance of meeting, or the limit that already may have been exceeded.

Only two years ago, many in the climate space called for radical emissions cuts to reach the goal. Today, the consensus among the business end of the climate community — asset managers, developers, business leaders and consulting giants — is that the goal is about to be in our rearview mirror, courtesy of the rise of AI. U.S. political headwinds and two more years of too little action? Just speed bumps on the way.

“The net-zero, 1.5-degree scenario is now a mathematical exercise. We could not get there from a policy perspective,” the leader of an energy analytics organization said. “We retain it because it’s important to look at it, but it’s not something that we were able to achieve other than [through] a mathematical formula.”

Move over Mitigation, the Age of Adaptation has Arrived

The other common climate term noticeably absent was mitigation. It used to be that mitigation was king and adaptation was a quitter’s word, one for those who didn’t believe we had the skills and will to pull humankind out of its climate nosedive.

Today, adaptation is a necessary evil needed to cope with the damage being done to the planet. It is not that everyone won’t do their best to mitigate, but it’s no longer enough.

“I’d like to avoid being seen as the pessimist in the room,” an in-house climate scientist at an analytics firm said. “We can turn this problem around and talk about the benefits of adaptation, or the benefit-to-cost ratio for adaptation. A recent study by JP Morgan has calculated that the reduced or avoided risk is between $2 and $43 per dollar invested in adaptation. It’s a huge opportunity.”

If you are building infrastructure today, the head of infrastructure at one of the world’s largest asset managers said, you might as well make it strong enough to withstand a future riddled with climate disasters. In the emerging markets, for example, he said it was “significantly cheaper to build something … to an international standard that will withstand the expected impacts of climate change that will come over the next couple of decades. We build to a certain standard that we hope we don’t need, but hurricanes have become stronger, fire risk is greater.”

Electrical transmission and distribution assets should be built with community vulnerability in mind, he said, so they not only could provide service during extreme climate events but also would not cause those disasters.

Business discussions about a post-1.5, adaptation-focused world were positioned as “pragmatic.” The word was used with a touch of guilt: not that they wanted to go gentle into that good night, but it seemed futile to rage, rage against the dying of the light. In an unwinnable war, pragmatists should focus on minimizing the damage … and perhaps even profiting from hardening the infrastructure ahead of the impending crises.

AI: Energy’s Frenemy

The rise of AI and the massive data centers that power it are the culprits behind the surrender of the target. But AI is seen as both a power-hungry enemy and an efficiency friend.

At Deloitte’s three-day Climate Week Horizons conference, data centers were cited as driving demand for electricity.

“There is no doubt that AI is creating the single largest [rise in] energy demand in decades,” said Martin Stansbury, principal and U.S. power, utilities and renewables risk and financial advisory leader at Deloitte. “And the real question is: Can we build it fast enough? Can we build it clean enough, and can it be resilient enough for the market?”

“By our analysis, we see about a five times growth in data center power demand by 2035,” Kate Hardin, executive director at Energy, Resources and Industrials Research, said at the Deloitte event. She said industrial electrification was the other major trend driving demand.

On the flip side, AI was cited as the potential hero for optimizing everything from buildings to ports … as well as the grid itself. While many speakers worried about how quickly generation could be built to service AI data centers or how the related demands on water supply would be met, others believed AI would unlock more energy savings than the data centers themselves would consume. Ultimately, AI will be a self-correcting challenge.

Speaking at Axios House, Tom Steyer, the billionaire co-executive chair of Galvanize Climate Solutions and former Democratic presidential candidate, talked about AI’s ability to make the grid substantially more efficient.

“If you look at renewables in this country, the big pain points are permitting and access to the grid,” he said, citing a discussion with a Californian investor-owned utility who told him it took 12 years to permit and build a new distribution line.

“The grid in California is 32% efficient. The grid in the United States is 42% efficient. We have the ability now, using real-time information and AI protocols, to completely change that number. If we can go from 32 to 64 [percent efficiency], we’ve rebuilt the grid,” Steyer said.

The chair of a major AI developer speaking at a different event had a similarly optimistic take: “The potential for efficiency and productivity of our existing infrastructure, using sensors, using artificial intelligence, is off the charts.” For example, data and sensors enable better use of the existing grid through dynamic load ratings. “When cable lines are cooler, they can get more current through them. Imagine if we could predict the weather, that would predict the temperature of the line, that would allow us to optimize how much current we put through. These are the sorts of things that are possible.”

AI’s potential to optimize generation assets would create further potential for gains, he said, increasing the output from existing renewables capacity. “Imagine that: 10 to 20% for effectively nothing.”

Efficiency Opportunities at Every Turn

It’s not just generation and the grid that could be net beneficiaries of the AI revolution, some argued, but every industry that depends on electricity will gain efficiency.

For example, the real estate sector is unmatched when it comes to climate risks and opportunities, said Lauren Pesa, partner and U.S. real estate sustainability leader at Deloitte. “Real estate is the largest asset class in the world, representing two-thirds of global wealth. First Street Foundation projects up to $1.47 trillion in real estate value in the U.S. could be lost in the next 30 years due to climate risk.”

Yet this massive asset class, a major consumer of the grid’s electricity, is ripe with opportunities to be both more resilient and more efficient. AI and machine learning are key for moving digital building management systems from set-and-forget to active optimization, said Ben Dwyer, SVP of global sales at Kode Labs, at the Deloitte event. “With the evolution of what’s going on with automation, AI and [machine learning], you’re essentially able to, over time, optimize the building to run as efficiently as possible.”

While optimizing energy use will be important, smart buildings’ ability to be more resilient during climate events such as Texas’ freeze and hurricanes will be critical, he said. “What smart buildings allow you to do is get ahead of these large events and ensure that your buildings are prepared for these things to happen, to ensure that not only the asset is protected, but the individuals that inhabit that building are protected,” Dwyer said.

Ports are another example. Beth Rooney, port director at the Port Authority of New York and New Jersey, told a Deloitte panel on aging infrastructure that the port is on the cusp of a transformation.

“Between today and 2050 … we’re going to double, if not triple, our cargo and passenger volume. I don’t have land to expand to, so we have to take the land that we have and use it more efficiently and more productively. And that’s where technology comes into play.”

The Port Authority, whose infrastructure includes aviation infrastructure, tunnels, bridges and rail terminals, was using technology ranging from underwater drones to inspect pilings (fun fact: cleaner water is degrading the port’s pilings faster) to GIS systems for more efficient maintenance. The ports were improving efficiency through four uses of technology, she said: predictive decision making, real-time risk management, enhanced visualization to see what’s going on across 3,000 acres, and hyper-connected logistics.

“We will not be able to handle the volume that is coming our way, not just in New York and New Jersey, but across the United States, if we don’t change the way we do business,” she said.

Wishful Thinking or Wondrous Optimization?

Overall, the consensus was that AI would save more energy than it consumed, leaving the world better off, even if there would be more fossil emissions to cope with during the transition. There was little talk about sequestration, despite the reluctant admission that fossil fuels would supply much of the AI-driven demand for now.

The head of an electrical components manufacturer warned against overestimating the amount of energy required by AI data centers. Last year, data centers consumed 1.3% of global electricity, so doubling by 2030 and doubling again by 2035 still would leave it consuming only about 5% of the world’s energy.

The question of how the markets would meet AI’s power demand brought out the list of usual supply suspects, most notably nuclear, with small modular reactors commercially viable and available within the next five years, and every color of hydrogen. AI still was in its relative infancy, and potential efficiencies, ranging from improved cooling technologies to quantum computing, were expected to lower energy demand as the technology matured. Similarly, value could be extracted from the waste heat by industry or district heating systems, where they exist.

The Waiting Game

Behind the barricades around the United Nations, President Donald Trump was ripping into climate goals when he wasn’t grumbling about halted escalators and broken teleprompters. However, business meetings mere blocks away seemed unconcerned that he labeled climate change a “con job” and wind and solar energy a “scam”. If anything, his comments evoked smirks and assurances that project development operates on timescales that stretch well beyond any single administration.

An industry analyst said, “No matter what the president says, 91% of new additions to the grid in the United States this year have been batteries and renewable energy. So, if you don’t think the transition is happening, you’re burying your head in the sand.”

And regardless of the president’s opinion, fiduciary duty required investment advisers to manage the $16 trillion of U.S. employee retirement funds prudently, a lawyer said. “What does that mean with respect to climate change? When we have all the scientific data, [it] means you have to actually be mindful that these risks are going to have a financial impact on the portfolio,” she said.

The dozens of speakers at events on and off the record agreed: Trump’s comments at the U.N., and his administration’s policies to date, were an annoyance, but not the driving factor for an industry that plans in decades. As the CEO of a major energy generation asset owner said, “There’s nothing I could build today that [will] have a payback period during the Trump administration. There’s nothing I could build today that’s going to have a payback period in the next three administrations.”

Practical Action, not Political Oration, is What Matters

Steyer said the administration’s policies may slow the energy transition, but couldn’t stop it: “Is [Trump] an existential threat to the world’s ability to make this transition and to solve climate change? No. We’re [the U.S.] 11% of emissions. Our emissions may go down slower, but … except for the year of COVID, we have never reduced emissions, and we’re supposed to reduce emissions 40% this decade,” Steyer said.

China’s action would far outweigh any U.S. inaction. “In the first half of 2025, Chinese emissions went down 1%. China is a third of global emissions. Between 2018 and 2024, 95% of new fossil fuel demand came from China. If China’s peaked, the world’s peaked.”

As Deloitte’s principal and sustainability leader, Geoff Tuff, put it: “The headline is, no matter what we hear in different parts of the world, and no matter what reports you may see, the energy transition continues, and it will continue. And it’s just the sheer reality. If you fast forward the clock 30 or 40 years, we will have a fundamentally different energy mix than we do today.”

A Week to End All Climate Weeks

Like the COP28 I attended in Dubai two years ago, it was good to experience Climate Week NYC, if only so I could check it off my list and never go again. I’m too old to shout over a techno beat at a climate-startup-investor schmoozefest, too jaded to hear one more high-octane luxury brand’s CSO congratulate their token coral-growing PR effort, and too tired to negotiate the subway sauna while running from one event to the next.

If I sound like a curmudgeon, I am. There’s no simple way to sort the hundreds of events across dozens of locations, no consistent way to apply for them, and no unifying calendar and map to navigate your choices. And while the coinciding event at the U.N. brings climate experts in from the most-impacted nations, it also drives hotel prices to astronomic heights, even by New York standards. Maybe they could build an AI agent to sort it all out.

But despite the stressors, this curmudgeon left happy to see that pragmatists are running the energy transition, tackling the big, hard challenges as they arise, undeterred by the political distractions happening a few blocks away.

Power Play Columnist Dej Knuckey is a climate and energy writer with decades of industry experience.