{"id":96727,"date":"2026-05-01T15:00:53","date_gmt":"2026-05-01T12:00:53","guid":{"rendered":"https:\/\/u1f987.com\/en\/?p=96727"},"modified":"2026-05-01T15:06:33","modified_gmt":"2026-05-01T12:06:33","slug":"when-intelligence-isnt-enough","status":"publish","type":"post","link":"https:\/\/u1f987.com\/en\/when-intelligence-isnt-enough\/","title":{"rendered":"When intelligence isn\u2019t enough"},"content":{"rendered":"

The artificial-intelligence industry has hit a barrier that money or new generations of chips cannot clear. A global grid crunch, component shortages and community resistance have turned the construction of data centres into one of tech\u2019s hardest logistical and political problems.<\/p>\n

How AI data centres differ from classical facilities, why the industry has grown more dependent on China, and how residents are voting out local officials to shield themselves from noise and environmental risks \u2014 all in ForkLog\u2019s new report.<\/p>\n

An architectural shift<\/h2>\n

Traditional data centres that powered the internet economy over the past two decades are fundamentally different from the architecture required for large language models.<\/p>\n

The classic data centre is built around the CPU<\/span> and draws on average 5\u201310 kW per rack, whereas GPU<\/span>-first AI racks consume roughly ten times more. Machine-learning racks with accelerators such as Nvidia\u2019s H100 or B200 require<\/a> 40\u2013120 kW each. The power-density gap changes the underlying physics of these sites.<\/p>\n

A cluster of tens of thousands of GPUs under peak load consumes as much electricity as a small industrial city. The snag is that distribution networks and substations are rarely designed for such abrupt, localised surges in demand.<\/p>\n

The appetites of AI\u2019s leaders have drained stocks of critical power-delivery gear: high\u2011voltage transformers, generators and batteries for uninterruptible power supplies. Manufacturing capacity in the US and Europe cannot keep up. As a result, lead times for industrial transformers, mostly sourced from China, have stretched<\/a> from one\u2013two years to three\u2013five.<\/p>\n

\"image\"
America\u2019s growing reliance on Chinese components runs counter to its political rhetoric. Source: Bloomberg.<\/figcaption><\/figure>\n

Another challenge is heat rejection. Air cooling cannot cope with AI-server density. The industry is being pushed toward liquid systems such as Direct-to-Chip<\/span> and immersion baths<\/span>. They require millions of litres of treated water and threaten regions, especially in arid climates.<\/p>\n

According to data cited<\/a> by Dutch researcher Alex de Vries-Gao, AI systems worldwide used about 765bn litres of water in 2025. To conserve resources, developers are refining designs. Instead of traditional cooling towers, where water evaporates into the atmosphere, new data centres are increasingly equipped with closed\u2011loop systems<\/a>. In them, water circulates through pipes, absorbs heat, is cooled in radiators and returns to servers with virtually no loss in volume. Adoption, however, is lagging far behind the pace of new builds.<\/p>\n

Almost at the Stargate<\/h2>\n

An unprecedented budget, top-level political backing and the mantle of the decade\u2019s defining AI alliance. At launch, Stargate<\/a> seemed to have it all \u2014 at least on paper.<\/p>\n

Stargate is a $500bn initiative that US president Donald Trump announced<\/a> in January 2025 as part of a national push to retain technological dominance. The OpenAI\u2013SoftBank\u2013Oracle joint venture was meant to drive a rapid build\u2011out of AI data\u2011centre infrastructure.<\/p>\n

A year on from the White House fanfare, the JV still lacked a full team and had not signed a single major construction deal in its own name.<\/p>\n

Markets turned skittish, too. JPMorgan Chase, slated to arrange $38bn of Stargate debt, ran into<\/a> investor doubts about the project\u2019s returns.<\/p>\n

OpenAI\u2019s Sam Altman and SoftBank founder Masayoshi Son disagreed on<\/a> fundamentals: where to build and who would be in control. From September to October 2025, Stargate\u2019s top brass shuttled to Tokyo for tough talks with Son but failed to settle who would own the platform for the flagship campus in Abilene, Texas.<\/p>\n

\"\u0430\u0431\u0438\u043b\u0438\u043d\"
A flagship site in Abilene spanning eight pads with a planned capacity of 2.1 GW. Source: Bloomberg<\/a>.<\/figcaption><\/figure>\n

OpenAI weighed going it alone, but lenders refused to provide funds without a clear path to profitability.<\/p>\n

Stargate scaled back<\/a> its aims and vacated a 900\u2011MW pad, keeping core Texas capacity with an eye to 1.2 GW later on. Partners shifted focus: in April 2026, developer Related Digital and Oracle raised $16bn in debt and equity to build a new mega\u2011data\u2011centre in Michigan for OpenAI\u2019s needs.<\/p>\n

Rivals seized the opening. In late March 2026 Microsoft snapped up<\/a> the vacant 900 MW, becoming Crusoe Energy\u2019s new partner to expand the Abilene campus. The upgrade should lift the site\u2019s total capacity to 2.1 GW by mid\u20112027, using Nvidia GPUs.<\/p>\n

People push back<\/h2>\n

Data centres are no longer viewed as unambiguous economic engines \u2014 they create few jobs once built, yet strain grids, consume water and generate constant noise.<\/p>\n

In April 2026 residents of Festus, Missouri, protested a $6bn data\u2011centre proposal. Locals secured the removal<\/a> of four of eight city\u2011council members and launched a petition to oust the rest, including the mayor.<\/p>\n

On April 9 residents sued the city, alleging it failed to give the public enough time to review the proposal before voting and made illegal zoning changes for the project. The suit also claims the city held private meetings about the plan instead of public ones.<\/p>\n

The approved scheme for an unnamed developer would cover 360 acres.<\/p>\n

Similar episodes have unfolded across America in recent months:<\/p>\n