Peter Schiff, chief of Euro Pacific Capital and longtime dollar crisis forecaster, has labeled the artificial intelligence infrastructure buildout a case of massive capital misallocation.
Speaking on his podcast episode “The Debt, the AI Bubble, and Strategy’s Liquidity Crisis… It’s All Connected,” Schiff argued that companies are spending roughly $1 trillion annually on data centers, GPUs, and computing infrastructure.
He warned that the equipment being purchased is “going to be obsolete 5 or 6 years, maybe, maybe quicker,” contrasting it with highways that last fifty years.
Microsoft (NASDAQ: MSFT) reported $30.88 billion in capital expenditure for its March quarter, representing an 84.39% increase year over year, illustrating the scale of spending across the sector.
Alphabet (NASDAQ: GOOGL) spent $35.67 billion in the same period, more than double the prior year, while Amazon (NASDAQ: AMZN) recorded $44.20 billion in a single quarter, with trailing free cash flow collapsing to $1.2 billion.
Meta (NASDAQ: META) raised its 2026 capital expenditure guidance to between $125 billion and $145 billion, an upward revision made mid-year, pushing the combined spending of these four companies toward $500 billion.
NVIDIA (NASDAQ: NVDA), the primary beneficiary of this infrastructure surge, posted Q1 FY27 revenue of $81.62 billion, up 85.23%, with data center networking alone growing 199%.
Jensen Huang described the moment as “the largest infrastructure expansion in human history,” and the market has responded by pricing NVIDIA at a $5.16 trillion market cap.
Schiff’s argument reaches beyond conventional concerns about stretched valuations, posing a fundamental question about the source and opportunity cost of this capital: “Where’s this trillion dollars coming from? What would all these companies have done with that trillion dollars if they weren’t using it to buy computer equipment?”
He pointed to mass layoffs and foregone investment in other areas as consequences, noting that consumer confidence sits at an all-time record low even as Wall Street celebrates the AI buildout.
Real-world enterprise experience is beginning to reflect the misallocation thesis, with Uber reportedly burning its entire 2026 AI budget by April after Claude Code spread across roughly 5,000 engineers faster than finance had modeled.
Uber’s CTO Praveen Neppalli Naga acknowledged the company was “back to the drawing board on its assumptions,” highlighting how AI coding tools are proving more expensive in practice than original capital expenditure models anticipated.
Stock price performance among the heaviest spenders has also raised questions, with Microsoft down 10% year-to-date and Meta down 2.3%, while Google and Amazon, where capital expenditure is more closely tied to measurable backlog and cloud growth, are up 22% and 21% respectively.
Vanguard’s 2026 outlook characterized AI enthusiasm as an “Economic upside, stock market downside” setup, while Goldman described the economy as masking weakness through “long-term transformative investments,” adding institutional weight to concerns Schiff has articulated more bluntly.
The bull case remains grounded in revenue momentum, with Microsoft CEO Satya Nadella reporting that the company’s AI business reached “an annual revenue run rate of $37 billion, up 123% year-over-year,” suggesting the spending is beginning to generate returns.
Polymarket assigns an 80% probability that Microsoft alone will be worth more than OpenAI and Anthropic combined by the end of 2026, reflecting market confidence that the infrastructure investment will ultimately prove justified.
The gap between Schiff’s depreciation timeline and the growth timelines projected by executives like Nadella and Huang represents the central tension this capital expenditure cycle will be forced to resolve.
