Broadcom CEO Hock Tan says AI growth is increasingly limited by capital, even as chip shortages ease

Broadcom CEO Hock Tan said on Monday that access to capital is likely to become the main constraint on artificial intelligence development, even as the industry works through earlier chip shortages. In remarks reported by Yahoo Finance, Tan suggested that the economics of scaling AI systems are shifting, with funding requirements becoming the gating factor for many projects.

The comment arrives as suppliers and customers move from the worst of the semiconductor tightness toward a more normalized supply environment. Broadcom’s framing implies that, once chip availability improves, the ability to pay for datacenter build-outs, networking, storage, and the supporting hardware and software layers can still limit how fast AI can scale.

Tan’s perspective points to a broader feature of the AI cycle: training and deploying AI models requires large, upfront investments, and those commitments depend on corporate budgets, investor appetite, and financing costs. When budgets tighten or capital becomes scarce, even companies that can secure chips may still slow expansion plans.

While Tan highlighted capital as the dominant constraint, the reported remarks did not specify which parts of the AI value chain would be most affected, such as compute procurement, custom silicon programs, or datacenter construction timelines. The post also did not lay out any quantitative view of AI spending, capex levels, or the pace of Broadcom’s own customer orders.

For Broadcom, the comment is relevant because the company supplies infrastructure-related products used across data centers, networking, and storage, areas that can experience strong demand when AI deployment ramps. If customers decide to slow or stagger investment due to financing constraints, that can influence order patterns across suppliers even when component availability is improving.

Broadcom’s sector context is also important. AI demand has been tightly linked to supply capacity in semiconductors and related components, but the scaling phase requires not just processors but the full stack around them. Tan’s message suggests the market may soon spend more time debating funding and returns on investment than resolving supply bottlenecks.

It remains unclear from the reported exchange what specific definition of “capital” Tan had in mind, or whether he was referring to corporate balance sheets, financing for capex projects, or constraints in broader market liquidity. The report also did not provide guidance on timing, and it did not connect the comments to any particular Broadcom financial metric or operating outlook.