Beware the “Investment”!
A crash-course in economic literacy as the AI bubble gets blown up with hot air and public funding
I recently learnt that a $20 ChatGPT subscription gets you access to about $700 worth of computing power.1 This inspired two ideas:
- WOW, I'd better turn my whole camera roll into Studio Ghibli-style remakes before this awesome opportunity runs out!
- RUN. NOW. Cancel your subscription. Escape with your brain.
Of course, the reason OpenAI and many other AI/LLM companies are willing to cover a significant discrepancy is because they are playing the long game. They see it as an investment, and imagine reaping untold financial rewards down the line.
Their strategy is that of a drug dealer offering a new buyer a free high, or that of a supermarket moving into a town and undercutting their own prices in order to drive out competition before jacking prices back up again. Their strategy is simple: create dependence.
It’s becoming easier by the day to imagine a world where people, businesses, and governments alike have embedded AI into their basic functioning to such an extent that they do not quite know how to operate without it. (Like, how did we function before the internet, right?)
AI to create, AI to design, AI to write, AI to think – a convenience becomes a need becomes an oppression. And like a frog in boiling water, the price we pay for it will go up and up. And we will pay it – if not in full out of our own individual pockets, then via our tax dollars, our environment, and our minds.
Already, governments are giving public land to data centres and helping out with their construction, providing them with public water and energy, letting them use infrastructure that we paid for and giving them tax breaks as incentives.2 Already, government-funded media companies are awarding contracts to foreign tech giants while laying off journalists.3
Why are we funding our own demise? Did someone say something about jobs and growth?
For decades, ‘jobs and growth’ has been the rallying call for all sorts of destructive developments: open cut mines, factory farms, nuclear facilities, soulless finance districts. For a moment, it was believable. Sacrifice the good, sweet things in life, and get prosperity instead.
But now, the development “vision” - represented by the fast-tracked publicly-funded data centre dynasty - is much more about destroying jobs than creating them. What is a data centre other than an enormous rectangular box that eats up water and energy for breakfast and our jobs for dinner? AI ‘development’ as an economic strategy is very obviously about laying off jobs for people and replacing them with jobs for robots and algorithms. It’s about using highly subsidised and actually very expensive inputs (energy, technology, rare earth minerals) to replace human time, creativity, skill and employment.
So, it makes my skin crawl when I hear blinkered politicians and tech lobbyists talk about the jobs created by data centres. Because the spattering of jobs they do create are surely the worst kinds: jobs for janitors who spend day after day inside dusting off the machines and pulling gunk out of the fans; IT people configuring data streams; engineers who test and monitor cooling pipes and backup diesel generators so that the computers don’t lapse for a second… god forbid there be a hiccup in their noble quest to turn my camera roll into Miyazaki rip-offs. These are by and large jobs people do not want to do.
- Private investment seeks private gain.
What does this nice word ‘investment’ actually mean? It means putting some money in to get more money out – pure and simple. Just like OpenAI footing the bill for my Studio Ghibli splurge on ChatGPT, ‘investment’ always has the end-goal of extracting wealth, even if it has a friendly front-end. With the AI investment boom shaping up to be the biggest in history,4 we are by definition also facing the prospect of a historic swindle. May the bubble pop before we’re all swallowed by it.
- Growth for the world’s richest corporations ≠ growth for the rest of us.
For generations, policymakers and the public alike have been fooled into thinking that economic growth has to look a certain way: bigger businesses + more foreign direct investment + more technology = prosperity. But remember in COVID how corporations like Amazon tripled their profits while the majority experienced a fast-onset economic crisis? Remember how, in the same decades that global trade and GDP boomed, the average middle-class family went from relying on a sole breadwinner to needing a double income?
The prosperity promise of “big is better” has always been a wolf dressed in sheepskin. With data centres the corporate global economy’s latest excuse for economic development, the gap between corporate growth and public prosperity is becoming harder to ignore.
- There are other possibilities for economic prosperity.
Attract and retain non-local business has long been the strategy employed by cities and states to fight for the scraps from the corporate economy. That is why governments offer enormous subsidies to attract warehouses, headquarters, factories, and now data centres.
Even on its own terms, this is an incredibly inefficient strategy of economic development. Research by economist Michael H. Shuman’s shows that, in traditional economic sectors like manufacturing and retail, economic incentives to smaller-scale, local businesses can create jobs for one-fifth to one-tenth the cost of recruiting a large external corporation.5 For example, Shuman points out how cities regularly offer hundreds of thousands of dollars to corporations in the form of tax breaks in return for a single job created. In the meanwhile, he states, “local economic development initiatives, such as entrepreneurship programs, deliver good jobs for $1-2,000.”6
My hunch is the wins of going small, local and low tech would be even more pronounced when compared to ‘investing’ in AI. At a crunch time in human history, there is plenty of work to be done – by humans, not by machines!
- Growing more food on diversified farms;
- getting that food to local markets;
- repairing infrastructure for regional processing and distribution;
- rebuilding national industries that had been offshored;
- caring for people;
- restoring watersheds;
- regenerating land that’s been deforested or denuded;
- building decentralised renewable energy systems;
- rebooting local newspapers, radio stations and media channels that tell relevant stories and share relevant information;
- reviving participatory democracy and decision-making systems.
So many things we need. So many opportunities for meaningful employment.
And yet, when it comes to “investing in our future”, our leaders are proposing giant thirsty warehouses of whirring lights and rare earth minerals and beepy noises and hot air.
Thankfully, I’m not the only one questioning this dumb vision of development. Communities from Spain to the USA to Australia are standing up to data centres in one of the more co-ordinated waves of resistance to infrastructure development we’ve seen in recent years.7 And, at time of writing, the call to put a moratorium on all new data centre development is growing by the day.
Talk to most ordinary people about AI, and the commonsense scepticism comes out in generous doses. I have sat through panels on AI in which industry leaders and technology researchers spout propaganda while audience members whisper to one another about their glaring doubts.
“No time for questions from the floor,” the MC informs us.
Now more than ever, we can’t leave the economy to the experts. The institutional thinking about “development” and “progress” has become increasingly narrow, increasingly technocratic, and increasingly disconnected from the broader human picture. Those who expound the benefits of AI have had their vision fenced in and blinkered by increasingly outlandish and outdated ideology. Their imagination has totally dried up. (Yes, probably aided by a $20 ChatGPT subscription.)
It’s time to trust our common sense, join the growing calls for a moratorium on all new data centre development, and get real about our future.
2 https://goodjobsfirst.org/cloudy-with-a-loss-of-spending-control-how-data-centers-are-endangering-state-budgets/(opens in new tab) / http://lasvegassun.com/news/2014/dec/30/lance-gilmans-43-million-tesla-payoff-has-critics-/(opens in new tab) /
https://www.businessinsider.com/data-centers-tax-subsidies-jobs-ohio-2025-5(opens in new tab)
5 Shuman, Michael H. The Small-Mart Revolution: How Local Businesses Are Beating the Global Competition. San Francisco: Berrett-Koehler, 2006.
6 https://michaelhshuman.com/vote-against-fake-economic-development/
7 Rauh, Reece & Simko (2026). ‘Local community opposition to data centers is clear and growing’. https://mreece13.github.io/files/dcenters.pdf(opens in new tab) and https://time.com/article/2026/07/22/community-backlash-ai-data-centers/(opens in new tab)
You can also read this essay on Medium(opens in new tab).
Illustration by Henry Coleman and Nicolas Coleman.
Middle photo by Geoffrey Moffett on Unsplash.
Stay connected
Insights, stories, and events from Local Futures — exploring localization as a pathway to ecological balance and community resilience.
Follow our blog
Thoughtful writing for a world beyond the global corporate economy — delivered to your inbox.
Author: Henry Coleman
Does the email system that you and I are using depend on data centres?
Im all for localisation but rely on email for communication.
Thanks Jay. I’m not sure about the technical question you ask – it may be the case that our emails need data centres.
Of course, we all have been steered into dependance on technologies – from the car to the internet. So the relevant discussion is not so much whether we as individuals are dependant on these technologies or not, but whether we allow policymakers to take us on another leap into increased dependance by building magnitudes more of the stuff. Localisation is more about steering the ship in a different direction – less about boycotting all technologies in our personal lives.