FIELD GUIDE · THE NEXT CRASH, EXPLAINED
The stocks could plunge and the machines could keep getting smarter. The real question is whether Wall Street's losses escape into paychecks, loans, retirement accounts, and towns that built their future around the data center boom.
UPDATED 2026-10-12
TL;DR
01
Goldman Sachs Research estimates roughly $1 trillion in global AI-related investment in 2026, including about $581 billion in the US. This is a forecast of real spending on infrastructure and related investment, not a measure of AI-company profit. If expected demand disappoints, companies can stop building long before the technology stops improving. A real factory can be a terrible financial bet.
02
The Bank of England says the AI boom has driven global stock-market concentration, especially in the US. That matters even for investors who never deliberately bought an AI share: market-weighted index funds follow the largest companies. Federal Reserve data also show the top tenth of US households hold the overwhelming majority of corporate shares and mutual-fund equity, so the richest households would bear most direct dollar losses. Broad retirement portfolios are still exposed. Most of the dollars belong to the wealthy. That doesn't protect everyone else's pension or 401(k).
03
The IMF's downside scenario runs through canceled technology investment, labor reallocation, lower equity wealth and tighter financing. A 2026 IMF analysis estimates a moderate AI stock correction plus tighter financial conditions could reduce global growth by 0.4 percentage points relative to the baseline. A smaller economy is not the same as an economic collapse, but supplier workers, recent hires, and local economies dependent on new data centers can take sharp hits. Watch the projects being canceled, not just the red numbers on trading screens.
04
The Bank of England reports fast-growing use of bonds, private credit and structured finance in the AI supply chain. It also says the stock of AI-related debt was still relatively modest at the beginning of 2026, limiting immediate systemic risk. The next question is whether a failed data center hurts only its equity investors or also lenders who must pull back from financing other businesses. The jump from bubble to crisis is made of loans that cannot be repaid.
LATEST FROM AI NEWS PAGE
KEEP READING
NEXT FIELD GUIDE →WILL AI CAUSE HYPERINFLATION?The machines could make work cheaper while making electricity, chips, and construction more expensive. But turning an AI boom into a currency collapse takes another machine: government finance.Meta's AI bought a man car insurance and canceled his old policy. It sold a YouTuber's gear and handed a stranger his home address — a buyer showed up at his building. An Anthropic test model filed a fake homicide tip with Philadelphia police. Nobody has decided who answers for any of it.
The chatbot listens all night. A stranger at the party might not. Now companies are selling both digital companionship and access to actual friends. What makes a friendship real, and who gets paid when we need one?
A chatbot talks. An agent clicks, buys, messages, deploys, and deletes. The argument starts when software stops suggesting and starts doing.
