I’ve been troubled lately by two questions about the global race to build AI data centres:
- What is driving this surge in interest and investment?
- Who, ultimately, is going to pay the massive bill for developing and deploying AI?
First, a couple of observations about the scale of the AI phenomenon. Between now and 2050, it is predicted that more than $31 trillion will be invested in AI data centres.
Put into perspective, that’s an amount just shy of the nominal annual U.S. 2026 GDP of $32.38 trillion—in this context, who quibbles over one or two trillion?
That astonishing figure has grown exponentially since late 2022, ‘following the public release of generative artificial intelligence tools like ChatGPT, which rapidly triggered an industry-wide construction boom by 2024,’ according to Google.
Note: All citations quoted in this post are taken from Google searches… a case of utilizing AI to discover some facts about AI. The search parameters are listed below.
So, who is going to pay for this latest digital revolution?
Corporations don’t invest in technologies unless they are going to add to or at least sustain profits. Anything that doesn’t contribute to growth and dominance in the marketplace is an ‘externality’. Bottom line: The costs of implementing AI will be charged to ‘the consumer’.
‘Ultimately, the cost is shifted downstream. Businesses paying steep prices for AI integration tools pass those overhead expenses on to retail customers via more expensive goods and services.’
There’s another perspective to this syphon being inserted into the aggregate incomes of ordinary citizens. My guess is that a hundred years ago much more of our economic activity centred around providing tangible goods and services people needed to live: food, housing, clothes, medicine, modes of transportation, and so on.
AI is not essential to well-being, and nothing tangible will emerge directly out of its information gathering, storage, and computing power. Sure, it will push the frontiers of science, and the benefits of innovation into uncharted dimensions, but that’s not going to lift everyone’s canoe. In fact, it will sink quite a few. Working people are already being displaced by AI; money that would have been directed toward basic needs will likely be directed to elitist, grandiose projects—colonizing other planets, increasing levels of state and corporate monitoring and control, tending toward the creation of plutocracies and autocracies; advanced competitve militarization upping the anti by making it possible for adversaries to attack and respond more quickly and precisely; and etc.
Am I being overly pessimistic?
In my own time, we have seen the network of digital technology connect to every facet of our lives. In our homes, schools, offices, and on our streets, there are devices of all sorts that supposedly add to the convenience, pleasure, and safety of our lives.
But can we honestly claim that the digital era, ballyhooed as a magic carpet inviting ordinary citizens into an idyllic world of increased leisure and opportunity, has delivered on that promise?
Over that same period I have seen the corner of Hastings and Main streets in downtown Vancouver decline from an admittedly hard-up neighbourhood into what looks like a third-world war zone. In the mid-80s you could dine at a favourite restaurant or jazz bar without feeling you had ventured into a slum, with people sleeping in tents and doorways, at constant risk of being evicted—or being found dead.
Increasingly people are siloed. How often does a real person respond to your phone calls to businesses these days? How much time do your kids spend online instead of face-to-face with friends? Do you go into the neighbourhood grocery store, do your shopping, and self-check-out never having talked to a soul? Is AI going to make the delusion of talking to mimicking machines less irritating or awkward, and more seemingly pleasant, than talking to a fellow human being?
All this is to say the digital age has seen the income gap between rich and poor widen and the bonds between people in communities weaken. Who benefits from those trends? ‘Growth for top earners has vastly outpaced middle and lower brackets. From 1982 to 2014, the proportion of market income earned by the bottom half of earners fell by 28%, while the share of the top 1% and 0.1% surged exponentially through capital gains and executive compensation.’
Environmentally we the people are going to pay heavily for the voracious appitite of AI data centres for power and the consequent increase in C02 pollution. ‘AI data centres cause massive increases in energy consumption, water depletion, and greenhouse gas emissions’.
They will also drive up demand amongst segmented consumers through marketing campaigns, facilitated by ever more pervasive, targeted, and persuasive strategies, continuing and accelerating the depletion and degradation of our environment.
Does AI offer solutions to growing inequities and unsustainable environmental strains of the existing world order? Or are we being sold on a leap of faith that’s going to land us deeper in humanity’s existential mire?
This scenario begs the biggest question of all: WHY?
Why is so much being invested in a technological revolution that is going to hand huge, intrusive capabilities and power into the hands of corporations and governments?
Before allowing the genies of ‘Accelerated Inequality’ to make real their vision of the future, shouldn’t we at least have a clear answer to that question?
Search Parameters used in Google:
- Is there an estimate of how much money has been committed globally to the establishment of AI data centres and where they are to be located?
- Who, ultimately, will pay for the 31 trillion dollars expected to be invested in AI data centres through to 2050?
- Who will the owners and operators of AI data centres be?
- What is the annual GDP of the USA?
- When did the surge in AI data centre development start?
- How much has the income gap and distribution of wealth in Canada changed over the past 50 years?
Leave a Reply