Do More Than Exist – But at What Cost?

1. Climate change, Scope 3 emissions + the hidden environmental footprint of AI

This summer, climate change no longer feels like a distant forecast.

It feels like a full blown heating air flow outside.

The World Meteorological Organization confirmed that 2024 was the warmest year in the 175-year observational record, at approximately 1.55°C above the pre-industrial average.

The following year was slightly cooler—supposedly the good news.

Yet Copernicus ranked 2025 as the third-warmest year on record, while the WMO confirmed that 2023, 2024 and 2025 were the three warmest years across all eight datasets it reviewed.

We cannot yet rank 2026 as a complete year, can we? In fact, still 5 months to form an overall observation. Somehow, though, the answer already feels quite clear. The signals are becoming incredibly difficult to ignore.

France experienced its 52nd recorded heatwave from 17 to 30 June 2026. Its 53rd began on 4 July. According to Météo-France’s analysis of heatwaves and climate change, two-thirds of the heatwaves recorded in France since 1947 have occurred during the twenty-first century.

In 2023, United Nations Secretary-General António Guterres declared that “the era of global boiling has arrived”.

Only three years ago, the expression sounded deliberately dramatic.

Now, in the heat of 2026, it feels considerably less rhetorical.

2. When the figures enter daily life

I have long been aware that extreme heat affects human health. Still, the recently reported figures were astonishing to read.

Preliminary data for 22–28 June 2026 showed 2,025 more all-cause deaths than during the previous week—an increase of nearly 30%.

This does not mean that 2,025 deaths have already been medically attributed to the heat. Final attribution takes time, and the initial dataset remains incomplete.

It is difficult not to suspect that the heatwave played a major role. But a reasonable suspicion is not the same as scientific attribution.

The distinction matters.

The wider pattern, however, is already established. Santé publique France estimated that more than 3,700 deaths were attributable to heat during summer 2024. The figure rose to more than 5,700 during summer 2025.

Facts and figures can seem dry and, at times, brutally abstract, until they enter daily life.

Temperatures above 30°C. Tropical nights. Fires. Disrupted transport. Altered opening hours.

Suddenly, climate risk is no longer a projection on a chart. It is how people sleep, work and move through their cities.

If attention is paid closely, The past becomes visible too. Old photographs in Paris museums are not climate datasets, and clothing alone proves nothing. Still, those images make the broader change feel strangely personal. By the way people dressed, the Europe pictured only twenty or thirty years ago hardly resembles the increasingly tropical summers we are experiencing now.

A photograph is not proof. But sometimes it connects the dots that data has already drawn.

3. When denial becomes a political hook

In 2016, The New Yorker revisited Donald Trump’s earlier claim that global warming had been “created by and for the Chinese”.

Climate denial may work as a political hook.

It does not work as an explanation.

The scientific evidence has continued to accumulate. The consequences have become harder to file away as somebody else’s future.

The question is no longer whether climate change deserves attention.

The question is whether governments, businesses and consumers are prepared to treat long-term planetary stability as seriously as short-term financial performance.

4. The difficult middle: Scope 3 emissions

There is some reason for optimism.

By the end of 2025, 9,764 companies had validated science-based targets through the Science Based Targets initiative—an increase of 40% in one year, according to the SBTi Trend Tracker 2025.

This matters because voluntary target-setting can move climate ambition from a communication exercise into operational planning.

The hard part is Scope 3.

Scope 3 includes emissions across a company’s value chain: purchased materials, transportation, suppliers, product use, disposal and other activities beyond its direct operations.

Under the current SBTi Corporate Near-Term Criteria, companies must include Scope 3 in their near-term targets when these emissions represent at least 40% of their combined Scope 1, 2 and 3 footprint.

Those targets must collectively cover at least 67% of reported and excluded Scope 3 emissions.

That threshold is not marginal.

CDP found that corporate supply-chain emissions were, on average, 26 times greater than emissions from direct operations. In many sectors, most of a company’s climate footprint sits outside its own buildings.

This is worrying because Scope 3 is both difficult to measure and difficult to influence.

Hence, crucially to solve the issues to the root, It should be more encouraging for companies to begin reporting the problem rather than hiding it. Transparency does not reduce emissions by itself, but reduction cannot begin without a credible map.

A value chain is only as resilient as its least visible link.

Or, put less corporately: no company is an island.

One business cannot decarbonise an entire value chain alone. Procurement teams need better supplier data. Suppliers need technical and financial support. Customers need lower-impact choices. Regulators need comparable standards. Finance teams need to connect carbon exposure with cost, risk and investment decisions.

Leadership also needs a clear vision, credible goals and climate KPIs that work across both the short and long term.

It is an all-hands-on-deck problem where collaboration beyond walls and borders more important than ever.

5. AI: doing more with less, but not without impact

I believe AI may become as transformative as the internet.

Used well, it helps people analyse more information, automate repetitive work and make better decisions with fewer resources.

But “doing more with less” must include the resources behind the screen.

In 2025, Google published a methodology for measuring AI inference. It estimated that a median Gemini Apps text prompt—not an individual token—used:

0.24 watt-hours of electricity;
0.03 grams of carbon dioxide equivalent; and
0.26 millilitres of water, or approximately five drops.

The disclosure was pioneering because comparable prompt-level data remains rare.

It was not the final answer.

The result depends on system boundaries, energy sources, location, cooling technology and whether indirect water consumption is included. A low impact per prompt can also become a substantial aggregate impact when prompts are generated by the billion.

According to the International Energy Agency’s analysis of energy and AI, global data-centre electricity demand increased by 17% in 2025. Electricity consumption from AI-focused data centres surged by 50%.

The IEA projects that total data-centre electricity consumption could roughly double by 2030.

Efficiency matters.

Scale matters too.

This raises a systems question: how many AI companies can measure and manage their environmental footprint while racing to become market leaders?

How many have the capacity or the incentive to optimise their models, use lower-carbon electricity, locate data centres responsibly and disclose water consumption before regulation forces them to?

Google’s methodology has attracted debate, as pioneering measurements often do. Still, publishing an imperfect methodology is more useful than maintaining perfect silence. It gives researchers, businesses and the public something concrete to examine, question and improve.

The companies that address these questions early may not be slowing innovation. They may be building the conditions for innovation to last.

6. Resilient profit or fragile profit?

The real choice is not between the planet and profit.

It is between resilient profit and fragile profit.

A United Nations General Assembly debate once described the environment and its resources as something we have “borrowed from future generations”.

We are not borrowing that future in theory.

We draw from it every day through carbon budgets, water systems, materials and energy.

The challenge is to make progress without quietly passing its full cost to the next generation.

If the trends of 2016 was “doing more than just exist” and doing more is as exciting as it sounds. Doing better with all costs in mind is more to the point now 2026.