IMF Models AI Bust at -1.1% Global GDP; Strategists Split on Risk

The IMF's financial stability report models a downside scenario in which a steep drop in AI spending leaves global GDP 1.1% below baseline by 2028, with the US the hardest-hit economy at 1.8% below baseline. The IMF flagged roughly $300 billion of circular financing between cloud giants and AI companies connected to $650 billion of debt. Jefferies strategist Chris Wood's base case is 'massive capital destruction in the US,' driven by cheaper Chinese open-source models capturing market share while the four largest US hyperscalers are expected to spend roughly $695 billion on capex in 2026 and roughly $870 billion in 2027. Chinese models processed 36.39 trillion tokens on OpenRouter in the week ending July 19, compared with 7.39 trillion for top US models. A Semafor columnist disputes that debt is the likely cause of any collapse, pointing instead to supply-chain bottlenecks across chips, memory, EUV machines, and raw materials as the binding constraint, and citing Anthropic's reported 85% gross margins as evidence of business soundness. The bull counterargument on capex is that cheaper tokens drive higher compute demand, keeping chip demand elevated regardless of which model wins.