SKN | Bank of America Challenges Wall Street’s AI Spending Fears With a Longer-Term Chip Market Outlook

Key Takeaways:

• Bank of America argues that slower AI infrastructure spending could extend the industry’s growth cycle, provided demand for paid computing remains resilient.
• The bank’s analysts believe slower supply growth could improve data-center utilization, support pricing and reduce the risk of excess capacity.
• BofA forecasts combined free-cash-flow deficits for eight hyperscalers and neocloud companies to rise from $128 billion in 2026 to $335 billion in 2027 before narrowing to $77 billion in 2028.

Bank of America is challenging the assumption that continued acceleration in artificial intelligence infrastructure spending is essential to sustaining the sector’s growth. In a research assessment following the U.S. midterm elections, the bank’s analysts argue that a more measured pace of data-center construction could ultimately support a longer-lasting AI investment cycle. The argument shifts attention from the volume of capital expenditure to the commercial returns generated by infrastructure already in operation.

Bank of America Sees Potential Benefits From Slower AI Supply Growth

The central argument in BofA’s analysis is that slower supply growth against resilient demand could improve the economics of data-center infrastructure. If new computing capacity expands too quickly, operators risk leaving servers underutilized and competing aggressively on rental prices. That would weaken the returns expected from substantial investments in chips, networking equipment and power infrastructure.

By contrast, a slower construction cycle could allow existing facilities to achieve higher utilization and stronger pricing. BofA’s analysts therefore suggest that lower peak earnings growth for infrastructure suppliers could be offset by a longer period of profitability. This is a significant distinction: the bank is not arguing that reduced spending is automatically positive, but that it may improve industry economics if customer demand remains strong.

BofA Highlights the Financial Pressure Behind the AI Buildout

The bank’s caution is supported by its assessment of infrastructure financing. BofA forecasts that combined free-cash-flow deficits among eight hyperscalers and neocloud companies could increase from $128 billion in 2026 to $335 billion in 2027, before narrowing to $77 billion in 2028.

These projections highlight the financial demands associated with the expansion of AI computing capacity. Even companies reporting rapid cloud growth must manage the gap between infrastructure investment and cash generated by operations. For BofA, the market may be recognizing capital-expenditure risks before demand indicators have materially deteriorated.

The bank also points to continued cloud demand. Recent reported revenue growth at Amazon Web Services, Microsoft Azure and Google Cloud suggests that demand for computing services remains substantial. Improving the productivity of existing graphics processing units can also increase revenue without requiring an equivalent increase in hardware purchases.

Political Changes Add Uncertainty, Not a Definitive End to AI Investment

The Democratic victory in both chambers of Congress introduces additional uncertainty around data-center development, particularly where local opposition, permitting and electricity infrastructure are already constraints. However, BofA’s central thesis is based on the relationship between computing supply and demand, rather than an assumption that political developments will immediately stop investment.

For investors, the distinction is important. Regulatory scrutiny could slow construction, but it does not necessarily eliminate demand for AI services. The commercial value of existing infrastructure may even increase if capacity additions become more disciplined.

What BofA’s Assessment Means for Long-Term Capital Allocation

Bank of America’s analysis encourages investors to look beyond headline capital-expenditure figures. The more important indicators are data-center utilization, cloud revenue growth, free-cash-flow generation and semiconductor orders. These measures will help establish whether infrastructure spending is producing sustainable commercial returns.

For internationally diversified wealth holders, the bank’s position offers a useful framework for assessing exposure to the AI investment cycle. A slower expansion does not necessarily signal structural weakness, but the thesis depends on demand remaining strong enough to absorb existing capacity. BofA’s longer-cycle argument will ultimately be tested by the ability of technology companies to convert substantial infrastructure commitments into durable cash flow and earnings.

For a confidential discussion regarding your cross-border banking structure and global wealth management arrangements, contact our senior advisory team.