In September, the people building the most powerful technology of our age asked to be slowed down.
Dario Amodei, who runs Anthropic, published an essay urging the industry to slow the pace at which it improves its models’ capabilities. Fellow AI leaders Sam Altman, Demis Hassabis and Elon Musk – not men who often see eye to eye – agreed about this one.
Within days, Senator Bernie Sanders and Representative Greg Casar introduced legislation to ban the development and deployment of artificial superintelligence.
President Trump was unimpressed, asking on his own social media network: “When, in the history of business, did anyone see the leaders of an industry call for regulation that, if strongly implemented, will drive them into oblivion and bankruptcy?”
To be fair, it is a reasonable question. What follows is my attempt at an answer.
As I see it, they called for more regulation because they can see something their own market cannot price. Each firm is locked in a race in which the rational move for any single competitor is to go faster, and the collective result of everyone going faster is a hazard none of them can avoid on their own. So the incentive governing each player runs against the outcome all of them need.
I have spent 30 years inside a different system with exactly the same problem. Finance is optimised for returns over the horizons its participants are paid to care about.
Climate change sits beyond those horizons. Its causes are collective, its gravest costs fall decades away and, once discounted into present value, those costs all but vanish. Faced with a risk hard to attribute to any one entity or quarter, the market simply assumes it away.
Unfortunately, the bill comes due all the same. Stranded fossil-fuel assets, measured as discounted future lost profits, are estimated at $1.4 trillion in the upstream oil-and-gas sector under the medium realignment scenario of a 2022 study in Nature Climate Change.
“Consider what it would do to policymaking to have thousands of banks, insurers and companies publicly advocating the same corrections at once”
Natural disasters caused around $224 billion of damage in 2025, of which $108 billion was insured, alongside approximately 17,200 deaths, according to Munich Re.
The longer the misalignment holds, the more abrupt the correction. A disorderly repricing could transmit losses across the financial system on a scale that would be difficult to contain.
A deeper reality
The parallel with AI is, I hope, clear. A market cannot correct a failure built into its own incentives, and the people inside it have both a licence and a reason to say so.
The Net Zero Asset Managers initiative understood this when it formed. Its commitment explicitly recognised that implementation depended on the policies and regulations adopted by governments and other public actors. The clause read as a disclaimer, but it pointed to a much deeper reality.
Mobilising the roughly $9 trillion a year needed by 2030 – and more than $10 trillion annually thereafter, estimates the Climate Policy Initiative – will require shifting the incentives that decide where private capital goes.
Only governments hold those levers: carbon pricing, standards, procurement and prudential rules that govern how much capital a bank must hold against a risk that is, for now, invisible on its balance sheet.
At least the AI industry has made its ask legible. It has told legislators, in plain terms, what it fears and what it wants done.
Finance has not done this. The issue has never really had the senior coordination or the collective buy-in that would push it to the top of the agenda.
We produce oceans of climate disclosure that tell an investor which company manages transition risk well and which might be stranded, and almost nothing that tells a government what to change, what is blocking it and what would move the capital.
This is the gap we must close, and fast.
Credible transition policies
Ask each company and financial institution to disclose, alongside its emissions, the policy measures that would make a credible transition pathway profitable in its sector – the carbon price, product standard, off-take agreement, fiscal measure and so on.
“Steering is government’s job. Ours is to draw the chart, and to broadcast what is becoming plainly obvious”
Aggregate thousands of those disclosures and you have a map of where scarce public money would draw in the most private investment.
Consider what it would do to policymaking to have thousands of banks, insurers and companies publicly advocating the same corrections at once. Consider, also, the coalitions that could form and the cover it would give nervous ministers.
There is, of course, a more tangible link between the two systems.
The data centres that serve and train these models are adding fast to electricity demand, and the same technology could just as easily sharpen our forecasting, balance our power systems and accelerate the science behind the transition.
Whether AI ends up a burden on the climate or a tool for good will be settled by the incentives we set around it, which is, once again, a question of governance rather than of engineering.
So back to Trump’s question. An industry’s leaders call for regulation when they can see – more clearly than anyone outside – that their own market is steering towards the rocks and that no competitor can turn the wheel alone.
Finance has been able to see its own iceberg for a long time. The difference is that we have preferred to steer around it in silence rather than recognise it and ask to be steered.
Naming a hazard, though, is not the same as avoiding one. Steering is government’s job. Ours is to draw the chart, and to broadcast what is becoming plainly obvious.
Steve Waygood is chief executive of the Finance Transition Centre.
