Top semiconductor executives from Intel and Broadcom warned that funding shortages could soon slow down the artificial intelligence boom. While demand for computing power remains massive, rising financing costs and unclear profit timelines threaten future infrastructure projects.
What happened
Speaking at the DBS Asia Future Forward Conference in Singapore, Broadcom CEO Hock Tan and Intel executive Lip-Bu Tan highlighted a looming financial bottleneck for artificial intelligence. Both leaders argued that securing funding might soon pose a greater challenge to the industry than physical shortages of silicon, memory chips, or electrical power.
Broadcom’s CEO noted that while hardware supply chain snarls have repeatedly emerged over the past two years, manufacturing capacity usually adapts over time. He projected that silicon wafer constraints could resolve by 2028 or 2029. In contrast, massive data center construction relies on steady debt and equity funding, making the cost of capitalThe expense a company incurs to obtain funding through debt borrowing or equity financing. a critical vulnerability.
Intel’s Tan underscored that corporate backers will eventually demand proof of profitability. As companies pour hundreds of billions of dollars into high-performance servers, questions linger over whether end-user software adoption will generate a sufficient ROIA metric that calculates the profitability of an investment compared to how much money was spent on it. to justify these massive outlays.
Why it’s important
Building out the digital backbone for advanced computing models requires immense financial resources. Technology giants have directed historic sums toward specialized chips, cooling systems, and real estate. However, higher bond yieldsThe percentage rate of interest paid on a debt security, which drives up corporate borrowing costs when it climbs. make corporate borrowing more expensive, potentially squeezing the profit margins on facilities that take several years to become operational.
Financial markets are also watching whether commercial usage keeps pace with infrastructure buildouts. Hardware makers have enjoyed record revenues during the initial wave of construction, but sustained order growth requires corporate customers to generate actual business earnings from their digital tools rather than relying on speculative budgets.
Modern hyperscale data centers can cost upwards of $1 billion each to build and equip with advanced artificial intelligence chips.
What could change
If borrowing rates stay elevated, technology firms may scale back capital expenditures, postponing new facility construction until credit conditions loosen. Alternatively, if enterprises rapidly integrate automated tools to achieve measurable cost savings, surge revenues could fund infrastructure growth directly from company cash flows.
A sudden pullback in data center spending could trigger revenue slowdowns for chipmakers if enterprise software fails to deliver quick financial returns.
What does this mean for investors?
Retail investors holding semiconductor equities like Broadcom (AVGO) and Intel (INTC), or broad index funds like the Invesco QQQ Trust (QQQ) and the VanEck Semiconductor ETF (SMH), should monitor corporate capital budgets closely. While near-term hardware backlogs remain solid through 2029, long-term equity valuations will depend on whether end users can translate technological investments into lasting profits.
Disclaimer: This post was created with the support of artificial intelligence and reviewed by our editorial team. It is provided for educational purposes only and does not constitute financial advice.