After years of funding Silicon Valley’s AI ambitions, investors now want more than grand visions. Meta CEO Mark Zuckerberg wrote in an op-ed that his company is committed to AI that would empower individuals and drive invention, predicting more jobs in the future and a more entrepreneurial economy. However, Meta’s AI spending will cost around $145 billion this year, and its stock fell by nearly 8%. Meta also missed earning expectations, racked up $2.4 billion in legal charges, and spent over $1 billion in restructuring and severance costs. Alphabet shares fell after revealing another increase in AI infrastructure spending, while shares in chip firms Nvidia, Samsung Electronics, and SK Hynix fell sharply. The share price of the Magnificent Seven (Nvidia, Microsoft, Alphabet, Amazon, Meta, Apple, and Tesla) tanked by $797 billion last week as AI skeptics dumped their stocks. The question is no longer whether AI will transform the economy, but whether companies spending hundreds of billions can afford the bill. The current boom bears similarity to overinvestment in railways prior to the panic of 1873, and AI has developed an unusual financial structure where the same tech companies invest in start-ups, sell them services, and buy stakes in one another, paralleling Victorian railway companies.