Apparently, we should have taken all the money being invested in AI and used it to remedy climate change instead.
At least that seems to be the argument coming from some of the more zealous corners of the climate movement.
It is an interesting comparison.
It is also rather difficult to understand what, exactly, they are comparing.
Because whatever one may think about artificial intelligence, large language models or the extraordinary expectations currently being attached to them, there is one rather important distinction between the AI economy and much of what has become the climate-industrial complex.
AI has customers.
Real ones.
Consumers want these services.
Businesses want them.
People are paying for them.
Companies are building products because there is an actual market waiting to buy those products.
That doesn’t mean the AI boom is rational.
Far from it.
The current expectations surrounding AI are almost certainly excessive. The projected size of the market, the assumptions about what these systems will eventually be capable of doing and the enormous investment pouring into the sector all have more than a whiff of fantasy about them.
Every technological revolution eventually develops its own mythology.
AI is no exception.
There will be spectacular successes.
There will also be spectacular failures, ridiculous valuations, stranded investments and companies whose entire business model turns out to have been little more than a PowerPoint presentation with a large number attached to it.
But beneath the bubble there is something rather important.
There is demand.
The market does not need to be invented.
It is already there.
Nobody has to force consumers to use AI.
Nobody has to pass a law requiring every company to buy a chatbot.
Nobody has to subsidise every interaction with an LLM in order to convince people that they want one.
The investment follows the expectation of demand.
Perhaps that expectation is wildly inflated.
But it is still an expectation of something that people demonstrably want.
That is a very different economic proposition from much of the climate-industrial complex.
Strip away the subsidies.
Remove the mandates.
Remove the regulatory preferences.
Remove the preferential permitting.
Remove the tax advantages.
Remove the government procurement programmes.
Remove the political pressure and the increasingly elaborate system of incentives and penalties designed to push consumers and businesses toward particular technologies.
Then see what remains.
That is the interesting number.
And I suspect it would be considerably smaller than the industry would like us to believe.
Because much of the activity associated with the climate economy is not driven primarily by voluntary market demand.
It is driven by policy.
The market is being constructed through legislation and regulation.
There is nothing inherently illegitimate about that.
Governments can decide that they want to subsidise particular technologies or accelerate particular forms of investment.
They have done it throughout history.
But we should then be intellectually honest about what we are looking at.
If an industry survives because governments mandate its products, subsidise its output, protect it from competition and penalise alternatives, that is not the same thing as an industry whose customers voluntarily demand its products at competitive prices.
Calling both of them “investment” does not make them economically equivalent.
And this is where the comparison with AI becomes almost comical.
The climate-industrial complex can point to enormous sums of money being invested and say: Look at the size of this market.
But how much of that market exists because people actually want to buy the product?
And how much exists because governments have decided that they should?
Those are very different questions.
If the subsidies and mandates disappeared tomorrow, much of the climate-related economic activity would not merely decline.
It would fall off a cliff.
That tells you something.
It tells you where the demand actually came from.
Perhaps the global climate economy is worth several trillion dollars.
Perhaps, depending on what is counted, the figure approaches ten trillion.
But a large economic number does not automatically prove the existence of a large underlying market.
A government can create enormous economic activity simply by compelling people to spend money.
The spending is real.
The buildings are real.
The factories are real.
The jobs are real.
The money is real.
But the demand can still be artificial.
AI has the opposite problem.
Its market may be wildly overvalued.
Its future capabilities may be wildly exaggerated.
Its projected revenues may prove embarrassingly optimistic.
But underneath the speculation sits something much harder to manufacture:
People actually want the thing.
So perhaps we should stop comparing the two as though they were simply competing destinations for capital.
They aren’t.
One is a speculative technology boom built on a genuine market.
The other is, to a significant extent, a policy-created market built around a political objective.
Apples and oranges would at least be botanically related.
This is closer to comparing apples with onions.
And wondering why they don’t taste the same.