Morgan Stanley has issued a stark warning about the financial strain facing the tech giants driving the AI revolution. The investment bank predicts that hyperscalers, including Microsoft, Amazon, Alphabet/Google, and Meta, along with neocloud providers, will face their most severe cash burn in 2027, with free cash flow turning negative and approaching $600 billion.
The forecast stems from an October 15, 2026 report highlighting AI-related capital expenditures. Morgan Stanley estimates total AI compute investment will hit approximately $1.58 trillion in 2027, a 14% increase over a baseline cash capex estimate of around $1.388 trillion. This surge is driven by lease financing obligations on top of direct cash spending, with AI investment projected at about $980 billion in 2026 and climbing to roughly $1.81 trillion in 2028.
The bank identifies 2027 as the peak year for the gap between spending and incoming cash, as infrastructure buildout accelerates faster than operating cash flow can keep pace. By mid-October 2026, hyperscalers had already issued approximately $255 billion in unsecured bonds, representing about 30% of projected full-year capex. High-rated firms like Amazon and Google are expected to maintain strong borrowing capacity, even with potential $400 billion in new debt in 2027, while lower-rated entities may face stricter financing constraints.
Nvidia is anticipated to benefit significantly from this wave of AI investment, with Morgan Stanley projecting the company could collect about $700 billion. The report also highlights the growing reliance on bond issuance as a core funding channel for AI infrastructure, signaling potential volatility in AI-related stocks and increased demand for semiconductors and server hardware.