Macro Pulse for October 25: Follow the Capital: A Dive into AI Buildout Financing

• Tech earnings last week highlighted growing investor scrutiny of AI-related capex. Despite better-than-expected earnings results, concerns over the cost of AI investment weighed on equities, with the Nasdaq falling over 3% on Thursday. Broader risk aversion drove demand for safe-haven assets, pushing the 10-year Treasury below 5% for the first time since September. We share our team’s detailed views on the outlook for AI capex and financing in this week’s note.
• The Federal Reserve is widely expected to leave rates unchanged at its meeting this week, allowing policymakers more time to assess economic data. Persistent inflation pressures caused and resilient economic activity support our expectation for one additional rate hike this year, at the Fed’s December meeting. 

Key Note

AI continues to be a defining theme of market performance in 2026 – sometimes for better, and sometimes for worse. From semiconductor volatility and the pipeline of AI-related IPOs to data center investment, we believe AI will remain a powerful market and economic catalyst, supporting corporate earnings, driving market performance, and bolstering economic growth in countries leading the AI buildout, particularly the United States.

But as AI spending accelerates and capital flows into the theme, the investment story is becoming more nuanced. What began largely as an equity story centered on semiconductors and hyperscalers is evolving into a broader capital formation cycle, with financing expanding into public and private credit, structured finance, and real assets, all while these growing investment commitments raise questions about whether AI adoption and monetization can generate sufficient returns to justify the capital being deployed.

In this note, we examine the expanding AI financing landscape, how AI has shaped markets and the economy in 2026, the outlook for translating investment into returns, and where we see the most compelling opportunities to gain exposure to this important theme.

The AI financing cycle continues expanding across public and private markets

While much of the initial AI capex was funded through hyperscalers' operating cash flow, an increasing share of financing is coming from external sources of capital. Capex among the five largest hyperscalers is projected to exceed $800 billion in 2026, up from $97 billion in 2020, while the U.S. AI infrastructure buildout could require more than $10 trillion in capital through 2032.¹

The AI ecosystem swim lanes by asset class

This is creating opportunities across public and private markets. Investment-grade hyperscaler issuance has led the way: more than $225 billion year to date, surpassing their combined issuance over the past four years.² But HY borrowing is also gaining momentum as neocloud providers (which supply computing capacity to hyperscalers and other AI companies) and data center developers increasingly turn to debt markets for financing

Private markets are also playing a growing role in financing the AI buildout, with estimates suggesting that private credit could fund roughly a quarter of global data center spending through 2028.³ Beyond direct lending, financing is expanding through project finance and securitization, which allows data center operators to issue bonds backed by their facilities and lease payments. Asset-based financing is also extending to chips and servers, with more than $60 billion in publicly identified GPU-backed financing commitments to date.⁴

The expansion of financing reflects the scale of the opportunity, but it also raises the stakes. As more capital flows into the AI ecosystem, the ability to translate AI demand into revenues becomes increasingly important.

AI’s economic promise does not guarantee investment returns

Investors are grappling with a harsh reality: AI has the potential to transform the economy and still produce disappointing returns for some investors. This distinction is becoming increasingly relevant as the companies funding much of the AI buildout face growing scrutiny over the returns on their investments. Hyperscaler equity performance has lagged in 2026 as returns on invested capital have fallen from recent highs as capex has grown faster than operating profits.

Even growing revenue commitments have yet to translate into a corresponding improvement in ROI – a phenomenon that is the result of a timing mismatch. Infrastructure investment is being made ahead of the revenue it is expected to generate, and the lag between capital deployment and monetization is likely to persist. A greater share of external financing (potentially reducing direct hyperscaler capex) relieves some pressure on hyperscaler free cash flow, but generates more sources of pressure for returns to be realized across financing streams.