George Noble, Managing Partner at Noble Capital Advisors and a former Fidelity Fund Manager, asserts that the current wave of artificial intelligence investment could become the biggest misallocation of capital in history. He suggests this could be 17 times larger than the dot-com bust, raising concerns about the substantial spending by tech executives on AI infrastructure.
Noble highlights that Big Tech's AI infrastructure spending is projected to reach $755 billion in 2026, up from $155 billion in 2022, and could exceed $1 trillion by 2027. Despite these massive investments, research from Wharton indicates that productivity gains would need to roughly triple for AI investments to break even at current trajectories, a benchmark that has not yet been met.
Noble also points to practices like circular vendor financing and "token maxing," which he believes obscure the true financial picture and inflate the apparent productivity of AI investments. He specifically notes NVIDIA's 770% growth in receivables, questioning whether demand is genuinely organic or artificially propped up by aggressive financing.
Adding to the concerns, Jefferies has warned that cheaper, open-source Chinese AI models could significantly challenge the US AI market. If these alternatives offer comparable results at a fraction of the cost, the competitive advantage of trillion-dollar infrastructure investments could diminish. Noble draws parallels to the dot-com era, when telecommunications companies spent approximately $65 billion on fiber optic cable that remained largely unused for years, suggesting AI could follow a similar, but much larger, script.
