Would cutting AI spending be more beneficial? Bank of America gave a counterintuitive judgment: although “unpopular” it may extend the boom cycle, chip stocks may usher in a longer profit runway

The Zhitong Finance App learned that recent market concerns have heated up, including the US midterm elections, concerns about rising AI infrastructure costs, and concerns about the sustainability of this spending spree. Bank of America said that while spending cuts of any kind may be an “unwelcome” sign for investors, it may actually prolong the spending cycle.

The bank's analysts put forward an inverse argument: if demand for monetized AI remains strong and production capacity growth slows down, then this may actually increase the supply capacity of computing power and delay the peak of the AI cycle.

“Slowing supply growth combined with remaining resilient demand may boost utilization, support pricing, and reduce the risk of excessive construction,” the bank's analysts wrote in a report sent to customers. “The result may be a downward trend in peak profit, but the profit runway is longer, and may result in higher valuation multiples for chipmakers that have been priced 'near the end of the cycle'.”

The bank is still optimistic about Nvidia (NVDA.US), AMD (AMD.US), Broadcom (AVGO.US), Maywell Technology (MRVL.US), Credo Technology (CRDO.US), Fanlin Group (LRCX.US), Kelei (KLAC.US), Micron Technology (MU.US), Adano (ADI.US), and Texas Instruments (TXN.US).

Further, the bank said that recently released earnings reports from hyperscale cloud vendors highlight the “resilience” of AI demand — the long-term contract backlog for these companies has reached a total of $2.5 trillion. In addition, Amazon (AMZN.US) AWS, Microsoft (MSFT.US) Azure, and Google (GOOGL.US) cloud platforms grew 39%, 49%, and 100%, respectively. While production capacity was limited, token processing increased more than tenfold year over year, and Hopper and Blackwell's rental prices rose 33% and 51%, respectively, during the year (SpaceX recently stated that current spot pricing supports a payback period of about one year).

Other notable data points include: Fortune 500 companies' spending on Amazon Bedrock increased more than 7 times year over year; Microsoft's number of customers with more than 50,000 Copilot seats increased 8 times (Copilot daily active users increased 50% month-on-month).

Investors have recently expressed concerns about the cash flow situation — involving free cash flow from eight US hyperscale cloud vendors and emerging cloud vendors. Free cash flow consumption is expected to expand from $128 billion in 2026 to $335 billion in 2027. However, analysts believe investors may “only see trees, not forests.”

“The 2027 calendar year is likely to mark a high point in cash consumption, and by the 2028 calendar year, the deficit will narrow to $77 billion, equivalent to 2% of sales, and an improvement of $258 billion,” the analyst added. “Funding remains the core controversy, but the ecosystem's free cash flow provides another potential release valve: Nvidia, Broadcom, and Micron are expected to total free cash flow of approximately $577 billion in the 2027 calendar year, and more than $800 billion in the 2028 calendar year (but still further exposed to supplier financing and circular trading risks).”

“As a result, the market appears to have factored in capital expenditure risk ahead of time before there was a rift in demand data,” the analyst continued. “In our view, if paid AI demand remains resilient, lower capital expenditure can reduce overconstruction, maintain scarcity, prolong the earnings per share cycle, and support higher valuation multiples.”