Abstract
How is the artificial intelligence (AI) investment boom being financed? We study the relationship between corporate cash holdings, investment, and financing during the AI era and compare it with previous technological transformations. We document two main findings. First, the association between current cash holdings and future investment is substan-tially stronger in the AI era than during the personal computer and internet eras, driven primarily by the relationship between cash holdings and R&D investment. Second, the AI investment boom is concentrated among a small group of large technology firms—the hyperscalers—whose cash-to-assets ratio declined from 35% in 2018 to 16% in 2026 while their investment-to-assets ratio nearly doubled. R&D led the initial investment expansion, while capital expenditures have driven its more-recent acceleration. The financing of this expansion differs sharply across firms. Data-center builders have combined a rapidly growing operating cash flow with lower payouts relative to internal resources and, increasingly, debt issuance to finance a large investment expansion. Semiconductor producers, in contrast, have experienced sufficiently strong growth in their operating cash flow to finance investment while increasing shareholder payouts and cash accumulation and reducing net debt. These findings show that the AI investment boom is associated not only with a reallocation of corporate cash toward investment, but also with sharply different financing needs across the AI supply chain.
Introduction
The rapid diffusion of generative artificial intelligence (AI) is transforming the scale, composition, and concentration of corporate investment. Developing and deploying AI systems requires substantial investment in R&D, semiconductors, computing capacity, and data centers. This investment boom follows several decades during which U.S. public firms—particularly the largest technology firms—accumulated substantial cash holdings. Since the release of ChatGPT in November 2022, however, capital investment by a few large firms has increased sharply while their cash holdings have declined. These trends raise the following three questions: Are firms using cash accumulated during previous decades to support the current investment boom? Is the relationship between cash holdings and subsequent investment stronger during the AI era than during earlier technological waves? How might firms finance continued investment as their cash holdings decrease? This article provides a descriptive account of these developments using firm-level data from U.S. public firms.
