Perspectives · Climate & Governance

The Hottest Summer on Record, and We Spent It Talking About AI. Heat Is a Project Condition.

I will say this plainly. I am tired. I just got back from a trip to the Middle East. I am home for a week, and then I head back there and on to Southeast Asia. Almost every meeting I sat in this summer came back to AI. Which tool to buy. Which jobs it will change. Whether the big models are moving too fast. I have opinions on all of it, and I have written about it. Meanwhile the heat did not let up, and it barely made the agenda.

The heat was global. The Copernicus Climate Change Service reported that June to August 2026 tied 2024 as the warmest northern summer on record for the planet. August tied July 2023 as the hottest month ever measured, at 1.65°C above pre-industrial levels. It was the first month above 1.5°C since November 2025. Western Europe had its hottest summer on record.¹ In the United States, NOAA reported the warmest summer for the lower 48 states in 132 years of records.²

The Global South took the hardest hits, in the places with the least room to absorb them. In India and Pakistan, the heat arrived in spring. From mid-April into May, many cities passed 46°C. World Weather Attribution found that climate change made the April heatwave about three times more likely. India set a record for electricity demand, and agricultural drought spread across more than one million square kilometers.³ Oman recorded 50.4°C in June.⁴ Human Rights Watch reported that migrant workers across the Gulf states still work outdoors through this heat without adequate protection.⁵ Tunisia had its hottest summer since its records began in 1950, with water shortages and power cuts.⁶ In Southeast Asia, a strong El Niño is drying out Indonesia. Forest and peatland fires are sending haze across the region, and a major reservoir that supplies Jakarta is dropping.⁷

When I raise this, someone usually says it is just El Niño. People say it as if it were an excuse. El Niño is a natural cycle, and it has always come and gone. Each one now starts from a warmer baseline, so the same cycle pushes temperatures to higher peaks. Copernicus attributed August's record to greenhouse gas warming and El Niño combined.¹ The South Asia heatwave in April came before this El Niño had formed, and World Weather Attribution still tied it to climate change.³ El Niño is also predictable. Earlier this year, NOAA put the odds of one forming between June and August at 62 percent.⁸ On a project, a risk that was forecast months ahead and left out of the plan is a planning failure.

These are the regions I am working in this fall. The people carrying the most risk in these places are outdoor workers, daily wage earners, and families in housing that holds heat. Very few of them were part of the AI conversations I sat in this summer.

I have worked in sustainability for more than 30 years, so none of this surprises me. It still makes me angry. The part that bothers me most is how little of the conversation the heat got.

Attention is a limited resource inside any organization. A leadership team can only carry so many priorities at once. A board has a fixed number of hours. A sponsor can only read so many briefings before the next decision is due. This year most of that attention went to AI. I understand why. AI is new, it moves fast, and a lot of money is moving with it. Heat feels like old news, even when it breaks records.

Heat is a project condition

Most projects happen in the physical world. Concrete cures differently in extreme heat. Crews cannot safely work full shifts when the heat index climbs, so many sites stop work in the afternoon. Power grids strain under cooling demand. Drought limits the water that construction, manufacturing, and agriculture depend on. Freight slows when rivers run low, and rail lines buckle. Each of these is a schedule risk, a cost risk, and a safety risk for the people doing the work. On many projects in South Asia, the Gulf, and Southeast Asia, the people doing that work are migrant and daily-wage laborers with the least protection and the least say.

Most project plans I review still treat weather as one line in the risk register, if at all. The entry is usually generic. It says "adverse weather," carries a medium rating, and holds a few days of contingency. That entry was written for a climate we no longer have. The baseline moved, but the plans did not.

AI and heat draw on the same systems

What frustrates me most is that AI and climate sit on separate agendas with separate owners. They rest on the same physical systems. AI runs in data centers. Data centers need electricity, and many of them use water for cooling. During a heat wave they draw on the same grid and the same water supply as towns, farms, hospitals, and project sites. Much of the new data center building is happening in the Gulf and Southeast Asia, the same regions that took this year's heat and drought. A company can announce an AI program and a climate target in the same annual report and never connect the two.

In my work on human adaptive capacity, I have argued that people and institutions can only absorb change at a certain rate. The same test applies here. Heat is drawing down the capacity of physical systems: water, power, soil, and the people who work outside. Fast AI adoption is drawing down the capacity of people to learn, judge, and adjust. Both compete for the same limited supply of leadership attention. Right now one of them is getting almost all of it.

I am changing my tone

I am over it, and I am done being polite about it. On this next trip, and in every meeting after it, I am going to challenge people on AI. Each time an AI program comes up, I will ask how it matches the organization's sustainability commitments. I want to see where the power and water come from. I want to see which targets the program supports and which ones it puts at risk.

If they are not matching the two, they are the problem. I will say that in the room.

What I am asking sponsors and PMOs to do

Keep working on AI. I write about it too. I am asking for three changes in how projects are approved and run.

I can ask this of sponsors because the standard of practice is now written down. The PMI® GPM® Guide to Responsible Project Sponsorship sets out what a sponsor is accountable for and how to carry that accountability through a project. No sponsor can say they do not know what best practice is. The guide is available at gpm.org/sponsor and pmi.org/learning/sustainability.

No one in the AI community ever asks me what should be built into their frameworks so I will openly offer my assistance if I am ever asked.  

First, update the impact baseline on every active project. Use local heat data from the last three to five years. Older averages understate the risk. Set work-rest thresholds based on wet-bulb globe temperature before the season starts, and build the lost hours into the schedule so crews are not pushed to make them up.

Second, put water and power in the AI business case. If a project adds AI capacity, the business case should show where the electricity and water come from and what happens to both during a heat wave. The P5 Standard already asks project teams to account for impacts on people, planet, prosperity, process, and product. It covers this without any changes.

Third, give heat an owner. In most organizations I work with, AI has a named executive and a steering group. Heat has a facilities manager and a safety memo. That difference in ownership explains a lot about where attention goes.

Next summer won't be cooler, and I expect the AI conversation to keep going. The people who approve projects should look at both on the same page, with the same care, before they sign.

¹ Copernicus Climate Change Service (C3S), ECMWF, "Surface air temperature for August 2026," September 10, 2026. climate.copernicus.eu

² NOAA National Centers for Environmental Information, national climate report for August and meteorological summer 2026, as reported by WeatherNation, September 13, 2026. weathernationtv.com

³ World Weather Attribution, "Climate change exposes hundreds of millions to longer and deadlier pre-monsoon heat in South Asia," May 2026; electricity demand and drought figures as summarized by Mongabay, May 2026. worldweatherattribution.org

⁴ Oman Civil Aviation Authority monthly weather statistics for June 2026, as reported by Gulf Insider. gulf-insider.com

⁵ Human Rights Watch, "Gulf States: Migrant Workers' Heat Protections Fall Short," June 23, 2026. hrw.org

⁶ Tunisia National Meteorological Institute, as reported by Al-Monitor, September 22, 2026. al-monitor.com

⁷ Associated Press, "Worsening El Niño threatens Asia with severe drought, potentially threatening global food security," PBS NewsHour, September 22, 2026. pbs.org

⁸ NOAA El Niño outlook for June to August 2026, as reported by The Southeast Asia Desk, "Southeast Asia braces for scorching heat, water crisis as 'Godzilla' El Niño looms," June 2026. thesoutheastasiadesk.com