Plans for a supposedly “renewable” multi-billion pound data centre project in Lanarkshire entirely misrepresented their actual environmental and economic impacts.
The Guardian‘s recent investigation reveals that the £8.2bn-valued data complex, touted as being energy self-sufficient, was unfeasible from the start. The centre and adjoining power-plant in Scotland’s central belt was supposed to be powered by on-site renewables and nuclear energy. In theory, anyway.
Built jointly by US corporation CoreWeave and Scottish data firm DataVita, the complex was supposed to power itself entirely by 2030. However, internal correspondence documents obtained through Freedom of Information (FOI) requests show that the site had an “issue” with “power provision” all along.
Government spokespeople and the private companies publicly promised that the Lanarkshire site would produce 1GW of “new energy infrastructure”. However, the two actors privately acknowledged that this entire promise was bogus and unattainable, the Guardian reports.
‘AI’ data centres built on lies and extraction
Before getting into this scandal, it’s worth reminding ourselves of some fundamentals about so-called ‘AI’:
- It’s not artificial. AI is built upon the large-scale corporate theft of human knowledge developed over millennia. It was trained over many years by underpaid semi-slave labourers in the developing world, often working in miserable conditions. It depends on endless amounts of natural resources like rare-earth minerals, copious amounts of clean water, and vast carbon-intensive energy sources.
- It’s not intelligent. We have largely been tricked into believing that complex large language models (LLMs) using plain English are human-like, if not superhuman, robotic entities. But AI lacks self-awareness, intrinsic motivation, and genuine comprehension. Instead of thinking, LLMs and machine learning algorithms perform complex, data-led statistical predictions. All they do is calculate the most likely next word, pixel, or decision based on past data. For this reason, they make countless foolish errors that no human brain with intelligence succeeding that of a three-year-old would make.
- It’s just a new technology. Humans are a bit like magpies. We love shiny, glossy, slick things. AI and LLMs are such things. Like many successful technologies in the past century — radar, GPS, cellular phones, velcro — AI emerged from the economic web surrounding the military-industrial complex. More than most technologies, it’s promising to devour our ecological sphere at exponential rates. These facts cannot be ignored. That doesn’t mean that it can’t be a net good in some areas, like vital healthcare or genuine public services. But only an economic model that prioritises these positive outcomes and which is readily willing to severely constrain tech companies will save us.
It’s all destructive extraction, corporate bluff, billionaire-funded marketing, and an impending economic bubble-pop. These are the fundamental lies at play, behind the more immediate lies in Lanarkshire.
CoreWeave + Gov = caught lying repeatedly
So the AI corporations are generally lying, and the British government knows this, but is seemingly lying about it too. But what about CoreWeave specifically? Well, in the same earlier Guardian report, they found:
CoreWeave’s press release said it would invest £1bn in the UK, and quoted the then technology secretary, Michelle Donelan, as saying this would bring “two new data centres to our shores.”
… Six months later, it announced that the two data centres were operational: one in London Docklands, and one in Crawley near Gatwick.
Planning records indicate that CoreWeave built no new data centres at either location during that time period. Neither did the two partners mentioned in its press release.
In fact, while CoreWeave’s – and the government’s – communications imply that physical buildings were built by suggesting the investment would bring “two new data centres to our shores”, this was misleading.
The same company is also facing multiple major class-action lawsuits (brought to civil court by more than one claimant) in the USA. On this legal battle, Yahoo Finance reports:
Multiple securities class action lawsuits have been filed against CoreWeave, Inc. (NasdaqGS:CRWV), alleging misleading statements about its capacity to meet customer demand and undisclosed reliance on a single third party data center supplier.
In essence, CoreWeave is yet another fibbing, self-inflating stock manipulation racket. It claims to be building data centres and funnelling investment when, in actuality, its economic model is akin to a pump-and-dump. It’s deceiving investors into investing and, apparently, gullible governments like ours into lucrative contracts.
CoreWeave’s partner in deception in Lanarkshire, DataVita, was also caught lying by the Guardian. It claimed to be generating the energy equivalent of supplying 800,000 Scottish homes, or nearly 15% of Scotland. All of this was supposedly “renewable.” But, in reality, the Guardian found on their website:
DataVita implies that all the energy sources are already in place, describing “energy parks” that are “directly connected” to its datacentres, delivering power that is equivalent to a small nuclear reactor.
But there is no evidence that DataVita currently has 1GW of private-wire renewable energy powering its data centres, or that it will have this in the near future.
So the question lingers: who’s being conned — the government, or us? Or is it both equally? Whatever the answer, we certainly deserve better than this incessant AI trolling.
UK joins the AI booster splurge regardless
Despite all this deception, as the Guardian‘s report writes:
The world’s biggest tech companies are now ploughing hundreds of billions of dollars into the buildout of AI data centres. Behind all this spending is the belief that AI will fundamentally transform the global economy, and that once it does, the AI data centres will pay for themselves.
Deceptive AI boosters have repeatedly caught the UK government with its trousers down already. Back in March, the Guardian revealed that the UK’s multi-billion pound drive to stay in the AI game is largely based on “phantom investments.” Their previous reporting states plainly:
… the money isn’t necessarily real, the data centres may not be new, the jobs are unaccounted for – and the supercomputer site 12 miles north of London is still a scaffolding yard.
The Labour-run UK government “rejected these assertions” in the Guardian‘s report, and claimed to have attracted over £100bn in AI-related investments since taking office. The government stated:
our AI sector [is] growing 23 times faster than the wider economy last year. That is delivering the jobs and opportunities hardworking people deserve.
This comment alone demonstrates the warped “growth, growth, growth” approach of neoliberal governments like ours. The so-called “AI sector” might be growing at 23x speed but it’s certainly not producing 23x jobs — we have an ongoing unemployment crisis raging. It’s not making Britain 23x better, socially or otherwise. If anything, almost 100% of that 2300% economic boom is benefiting far-away tech oligarchs only.
Featured image via the Canary
