The Brick Lane data centre is Labour’s ‘market Stalinism’ in action

Plans to build data centres have been controversial across the UK, but the latest example comes from Brick Lane.

Housing Secretary Angela Rayner approved a £500 million redevelopment of the Truman Brewery on Brick Lane, which will include a data centre. This is despite unanimous opposition to the development from Tower Hamlets council, the local mayor and campaigners.

In a borough with over 30,000 households on the social housing waiting list, the council wanted 800 homes on the same site instead. The government’s decision letter gives that plan “no more than limited weight.” Meanwhile, it gives the data centre “significant weight” for its economic benefits in meeting “the need for data centre capacity,” but never states how many jobs it will create.

The Save Brick Lane coalition argues the site should be housing-led, warning the development could push up rents, displace long-standing residents, and hollow out Brick Lane’s community. With unanimous opposition by locals and 30,000 people waiting for homes, the government pushing ahead with this data centre shows in whose interests the government is willing to make tough decisions.

Computer says no

The late great theorist Mark Fisher might refer to the government’s behaviour in this case as “market Stalinism.” Fisher identifies a system that opposes centralized, bureaucratic and authoritarian Stalinist planning, but, in fact, reproduces it via capitalism.

Under this system, the pursuit of material progress for everyday people is overlooked in favour of hitting targets for profits or growth, standing in for Stalin’s five-year plans, so that the metrics become the ends in and of themselves. That means branding achievements become more important than achieving any sort of material benefit for ordinary people.

A Guardian investigation showed how much market Stalinism is at play in the UK’s AI strategy. CoreWeave’s announced £1 billion UK “investment” turned out to be renting space in two existing datacentres and plugging in chips it had already bought without building anything new. Nscale’s £2 billion “supercomputer” – that the government said would be completed by the end of 2026 – was still just a scaffolding yard when reporters visited in February.

Ordinarily, market Stalinism at least performs the audit; here, the Department for Science, Innovation and Technology didn’t even manage that, admitting it wasn’t:

playing an active role in auditing these commitments.

Economist Cecilia Rikap speaks of ‘phantom investments’ whereby:

big tech companies artificially inflate datacentres’ job creation and economic impact to please governments [who] are desperate to claim they are making the economy grow.

Britain’s AI roadmap

Britain’s AI roadmap was supposed to have accounted responsibly for its own carbon costs, yet the Department for Science, Innovation and Technology had underestimated the emissions from AI data centres by roughly a hundredfold. A since-deleted estimate of 0.142 million tonnes a year was  revised to a range of up to 123 million tonnes over the next decade.

This estimation was only corrected after the campaign group Foxglove and the outlet Carbon Brief checked the working themselves. The correction landed a week after Kendall urged the public to “embrace AI” after announcing the first £500 million tranche of a new government AI fund. The ministries love the announcements, but how often do they check it against the world?

Brick Lane fits that pattern. A data centre approved for “significant” benefits nobody has to specify, while the quantifiable loss – more homes in an area needing homes – is overshadowed in the pursuit of growth. The case is not isolated. Data centres were designated critical national infrastructure, and planning rules were rewritten to give data centre applications more weight. Ministers even proposed folding projects into the fast-track regime for nationally significant infrastructure. That urgency might relate to the financial story behind AI.

Betting the house

AI-linked firms have climbed in value by $27 trillion in 3 years, built on what Goldman Sachs analysts called “Panglossian” profit assumptions</a. The Bank of England’s Andrew Bailey has warned that an AI bubble bursting could shrink UK GDP by 2.2%, while the IMF’s Kristalina Georgieva has compared today’s AI boom to “the bullishness about the internet 25 years ago”, which resulted in the dot-com crash.

Local opposition to data centres could play a part in bursting the AI bubble. Each local area that refuses to have a data centre, to not participate in building out AI’s physical infrastructure, contributes to a mechanism that could trigger a correction of the Panglossian profit assumptions. When national governments view local opposition to data centres, they are not just assessing local planning disputes, but deciding whether an overleveraged bet pays off.

The way in which governments make decisions is at the heart of the issue. The case of the Truman Brewery was not simply about creative differences, but also testing whether the alternative Tower Hamlets wanted – 800 homes instead of a data centre – was ever seriously weighed at all.

That phrasing of the government’s decision letter does the same work as Thatcher’s “there is no alternative”. It does not seek to defeat the opposing case, but rules it outside the range of things a government needs to weigh at all, granting it an unbearable lightness. All that is solid then melts, as the government looks at the human need for housing as an obstacle to a financial story it has already decided must come true, bending reality to the bubble.

Featured image via BBC

By Hugo Harvey

Source link