Ideology at the Top, Infrastructure at the Bottom

Artificial intelligence has become one of the few issues in American politics that still enjoys bipartisan support. Yet that consensus increasingly resembles a check written on trust—its terms vague, its returns uncertain, and its demands ever growing. Amid lofty declarations about "the future of humanity," the largest tech companies are asking for very tangible things: land, electricity, infrastructure, and regulatory shelter. All of it in service of tools whose public benefit remains contested.

Almost immediately after taking office, Donald Trump signed an executive order aimed at reinforcing U.S. leadership in the field. His administration announced support for a $500 billion project: a consortium of OpenAI, SoftBank, and Oracle plans to build a network of data centers and infrastructure to scale language models like ChatGPT.

Democrats, too, have embraced the cause. Senator Chuck Schumer declared: "If America loses to China in the AI race, we lose everywhere—economically, militarily, scientifically, and educationally." On Thursday, a Senate committee held a hearing titled "Winning the Race for Artificial Intelligence." Among the invited speakers were OpenAI CEO Sam Altman and Microsoft president Brad Smith. The discussion centered on "regulatory barriers" and the need for "U.S. hegemony in the industrial revolution of the 21st century"—with China cast as the main rival.

Behind the soaring rhetoric about AI’s world-shaping potential, there was a striking lack of specifics. Participants spoke of so-called AGI—artificial general intelligence—without defining what exactly they meant. In written testimony, Sam Altman admitted the term was "loosely defined," and suggested it be understood as "a system capable of solving increasingly complex tasks at a human level." He called AI "the most powerful tool ever created," capable of delivering "an almost unimaginably bright future"—if, that is, the American model prevails, one rooted in "democratic values like freedom and transparency." Brad Smith echoed the sentiment, calling AI "the most helpful tool for humans ever invented."

Beneath the lofty tone, however, lies a more grounded objective: securing government support. To win the race, Smith emphasized, the U.S. would need "massive data center infrastructure"—which means federal assistance in expanding and upgrading the electrical grid. In 2023, OpenAI spent $1.76 million on lobbying—seven times more than the year before. Its demands include: a ban on restrictions from individual states, export controls to prevent technology leakage to China, loosened intellectual property rules, accelerated grid construction, and access to government databases.

Congress welcomed these requests with enthusiasm—no one asked uncomfortable questions. Such as: why are companies seeking large-scale public subsidies still unable to turn AI into sustainable profits? OpenAI ended last year with $5 billion in losses. Microsoft has already shelved several U.S. and European projects totaling 2 gigawatts—citing overestimated demand.

Meanwhile, real-world applications of language models remain far from revolutionary. Social media is flooded with Ghibli-style images and AI-generated captions. Email clients offer condensed message summaries. As Harvard Business Review notes, the most common use is that of a virtual companion—for advice, emotional support, or declarations of affection. In universities, cheating has gone mainstream. According to New York Magazine, students are using ChatGPT to write essays and fill out applications. "We’re going to have a whole generation of graduates who effectively don’t know how to write," said one professor from California.

The social consequences remain absent from the policy agenda. Yet the damage is already visible: cognitive decline, digital noise, and diminished empathy. If AI is truly so vital to America’s future, why is it making everyday life in America feel increasingly unlivable?