Full Breakdown
AI Boom Faces a New Bottleneck: Capital Availability
By Drooid · · How we work
Executive Perspectives on the Emerging Constraint
At the DBS Asia Future Forward Conference in Singapore, Broadcom chief executive Hock Tan warned that “capital” could become the biggest constraint on the artificial-intelligence surge, eclipsing physical shortages of semiconductors, memory chips or electricity. Lip-Bu Tan, chief executive of Intel, echoed the concern, noting that rising interest rates are inflating the cost of capital and that investors are increasingly demanding clear returns on AI-related spending. Both executives stressed that while demand for compute infrastructure remains robust, financing challenges could shape the sector’s trajectory beyond the next two to three years.
Industry Context and Recent Supply Bottlenecks
Over the past two years the AI ecosystem has repeatedly confronted shortages—first in chips, then in memory and power—and has generally managed to expand capacity after each bottleneck. Hock Tan suggested that memory constraints may ease within a year and that wafer shortages could improve by 2028-2029, with power availability also expected to rise as new infrastructure comes online. Nonetheless, the executives argued that capital constraints are less amenable to engineering solutions.
Implications for the AI Infrastructure Market
Higher bond yields are raising financing costs for data-center construction and chip procurement, making long-term projects harder to justify when debt service climbs. The executives warned that the sector’s growth beyond the near term will hinge on whether businesses can integrate AI into operations and achieve tangible economic benefits. Lip-Bu Tan also highlighted competitive pressure from Chinese models—such as DeepSeek, Kimi and GLM—whose lower token costs and rapid adoption could temper demand for the most expensive systems.