We all know that something changed in America in 2016, even if we\u2019re still struggling to sort out what happens next. President Donald Trump broke through the \u201cblue wall\u201d of Pennsylvania, Wisconsin and Michigan as rust-belt voters agreed with his campaign messaging that they\u2019d been \u201cripped off\u201d and \u201ctaken advantage of\u201d and the system had been rigged against them. Decades of promises from politicians in both parties who negotiated free-trade deals hadn\u2019t panned out, and groundbreaking work from economists David Autor, David Dorn, and Gordon Hanson gave the era a name: \u201cthe China shock.\u201d<\/p>\n
Even JPMorgan CEO Jamie Dimon recently admitted<\/a>, in an onstage appearance with Anthropic CEO Dario Amodei, that the government\u2019s post-NAFTA promise to reskill manufacturing workers \u201cdidn\u2019t work,\u201d saying \u201cit wasn\u2019t set up right.\u201d But Dimon expressed optimism that government and business can work together better this time, somehow.<\/p>\n Count Bhaskar Chakravorti, dean of global business at Tufts University, as a skeptic. He just mapped out the American AI Jobs Risk Index<\/a>, a model tracking the geography of jobs most vulnerable to AI automation across 784 occupations and he\u2019s certain that the \u201cRust Belt\u201d of the China shock era is getting a \u201cwired belt\u201d in the AI era.\u00a0<\/p>\n According to Chakravorti\u2019s index, 9.3 million American jobs are vulnerable to AI automation, amounting to a projected $200 billion in lost income. But in an extreme scenario where AI is able to replace a larger share of labor, that figure rises to $1.5 trillion. Most of those are concentrated in just a handful of metros, according to Chakravorti.<\/p>\n \u201cThere are 14 knowledge-driven metros that range from the whole San Jose area to the Raleigh-Durham area to big cities like New York or Seattle or Boston,\u201d he told Fortune. \u201cThey face 3.6 times the job loss and over five times the income loss of the traditional sort of manufacturing and tasks.\u201d<\/p>\n A quarter of a century after companies started offshoring manufacturing jobs to China, AI stands to have a similar impact, albeit in office buildings across corporate America rather than in factories. While the national unemployment rate stayed relatively low outside of major downturns, it ran persistently higher in manufacturing-heavy cities like Detroit<\/a>.\u00a0<\/p>\n \u201cIt\u2019s heavily concentrated either in the coasts or in the knowledge belts in [and] around universities,\u201d Chakravorti said. \u201cThose are the kind of the major areas that are likely to see displacement.\u201d<\/p>\n AI-linked layoffs are rising. But the broader job market is steady<\/strong><\/p>\n With the possibility of AI automating many roles in the white-collar world in highly-concentrated urban and suburban areas, those communities could hollow out in the same way. The decline of Detroit took several decades. Some AI executives, like Microsoft<\/a> AI chief Mustafa Suleyman, think half of entry-level white-collar roles will be eliminated<\/a> within a year-and-a-half.<\/p>\n The current AI layoff landscape isn\u2019t too dire\u2014yet. Outplacement firm Challenger, Gray and Christmas<\/a> recently reported that 49,1235 layoffs have been associated with AI automation so far this year, compared to about 55,000 in all of 2025. However, a portion of those layoffs stem from tech firms like Meta<\/a> and Microsoft, which have cut workers to free up cash for increased investment in AI infrastructure. Tech unemployment\u2014the sector where most AI layoffs have taken place\u2014ticked up to 3.8% last month, but still remains below the overall unemployment rate of 4.3%.<\/p>\n Why an impact similar to the China Shock may not be a bad thing<\/strong><\/p>\n However, not everyone draws pessimistic conclusions from the parallels between today and the China shock. Apollo chief economist Torsten Slok said in a note last week that the similarities between the AI shock and the China shock<\/a> is actually a good thing. His logic follows that cheaper intermediary goods from China actually helped to boost manufacturing productivity. That resulted in a 50% increase in real manufacturing value from 2001 to 2024. He foresees the same thing happening with AI.<\/p>\n \u201cIf history is any guide, the gains will be substantial,\u201d he said. \u201cJust as cheaper Chinese inputs helped U.S. businesses grow and hire, AI is already accelerating business formation and productivity gains across the economy.\u201d<\/p>\n Slok also noted that AI is already spurring new business formation and driving productivity gains.<\/p>\n Still, Chakravorti said the AI shock could have a profound impact on the knowledge workers in the country\u2019s biggest urban hubs. And that impact could reverberate across the political landscape. He said that in the same way the Rust Belt helped to elect President Donald Trump, so too can what he calls the \u201cWired Belt<\/a>,\u201d the areas in which knowledge workers stand to be displaced, can build a formidable political movement.<\/p>\n