JPMorgan recently dropped a research report stating that it's bumping up the AI total capital expenditure forecast before 2030 from $5.1 trillion to $5.5 trillion, with corresponding debt financing also raised to $4.1 trillion. The report highlights that the surge in AI demand combined with a persistent shortage of computing power is driving this spending acceleration. Google alone has hit an AI token processing volume of 32 trillion times in a month, while companies like Microsoft and Uber have blown through their entire annual AI budgets in just a few months this year. Among the major cloud providers, the four US supercomputing giants (Google, Amazon, Microsoft, Meta) are guiding for a combined capital expenditure of about $700 to $725 billion by 2026, marking a roughly 75% year-over-year increase, with expectations to surpass $1.1 trillion in 2027. JPMorgan anticipates that their total operating cash flow will exceed $900 billion by 2027, but it still won’t be enough to cover the massive spending plans, making debt and equity financing the norm. On the chip front, JPMorgan expects Broadcom to achieve over $150 billion in AI-related revenue by 2027 (including ASIC/XPU and AI networks), with management disclosing that the backlog for 2027 exceeds $100 billion, far exceeding the bank's expected $60 billion for 2026, which is seen as a conservative estimate. The AI chip financing platform (AI XPV) formed by Broadcom in partnership with Apollo and Blackstone is also reported as a landmark structural innovation in GPU financing. In terms of financing structure, the five supercomputing companies have completed about $240 billion in external financing year-to-date, with the high-rated bond market still being the main channel. It’s projected that in the next five years, the issuance of high-rated bonds related to AI will reach $2.1 trillion. On the power front, JPMorgan has raised its data center capacity growth forecast from 122GW to 138GW, but power remains the most critical bottleneck, with supercomputing vendors actively exploring alternative solutions like self-built power sources.