AI boom: How AI spending is impacting debt markets
This article is the first in a series from Mariner’s investment team that explores how the artificial intelligence (AI) boom is impacting different asset classes. We’ll look at how debt, equity and private investments are being driven by AI and what it means for your portfolios.
By Luca Vujovic, CFA, Senior Portfolio Manager, Fixed Income and Brian Baker, CFA, Director, Investment Strategy
Spending spree
To understand how AI is impacting debt markets, you must first understand the massive amount of spending that is taking place to support the buildout of AI infrastructure. Take the so-called hyperscalers, which include tech giants Alphabet, Amazon, Meta Platforms, Microsoft and Oracle. You only have to go back three years to 2023 to see their combined capital expenditures of about $145 billion.
Fast forward to 2026, when these hyperscalers are estimated to spend a combined $750-$825 billion, according to estimates from Bloomberg Intelligence, which represents a nearly sixfold increase from 2023. The spending is projected to rise further and top $1 trillion each year from 2027 to 2032.
Historically, most of these businesses were extremely cash generative, with modest levels of capital spending, which allowed them to generate copious amounts of free cash flow for shareholders. But in order to meet their enormous AI spending needs, hyperscalers are increasingly being forced to tap debt markets.
Borrowing boom
In general, mega-cap tech companies have operated asset-light businesses that generated far more cash than could be deployed internally, making them infrequent and opportunistic visitors to the debt markets. Today, the capital intensity of the AI infrastructure buildout has outpaced even their enormous levels of operating cash flows, transforming some of the world’s most cash-rich companies into the primary source of new supply in the U.S. investment grade bond market.
Throughout 2024 and the first three quarters of 2025, hyperscaler bond issuance was sporadic and modest. That dynamic changed dramatically in late 2025, when Meta, Alphabet and Amazon issued a combined $70 billion in debt in a single month. The issuance has only accelerated in 2026, with well over $130 billion raised by hyperscalers in the U.S. investment grade market as of mid-October , with the number climbing by the week. The number is over $212 billion globally, when including non-USD notes.
Furthermore, debt reported on balance sheets tells only part of the story. Behind these figures sits a massive number of additional obligations and commitments, including recognized data center lease liabilities and commitments for leases that have not yet commenced.
To be sure, this debt accumulation at the hyperscalers has occurred alongside, not in the absence of, massive operating cash flow generation. Today, these are not distressed borrowers, and their operating income covers current interest expense by a wide margin. But the pace and scale of balance sheet expansion is a structural shift that demands attention, particularly as spending guidance continues to be revised higher and free cash flow compresses or turns negative.
ROI
When analyzing a capital cycle of this magnitude, the central question for bondholders is whether future cash flows will justify the debt being issued today. Total generative AI revenue across hardware, software, services and advertising is projected to expand from roughly $380 billion in 2025 to over $2.3 trillion by 2032, according to Bloomberg estimates. This represents a compound annual growth rate of about 30% and would see AI grow from 6% to 22% of all global IT spending.
As data centers are completed and the potential returns from AI scale across existing infrastructure, capital spending growth should moderate while operating cash flows compound, driving a strong recovery in hyperscaler free cash flow generation in 2029 and 2030. For bondholders, these potential high-margin cash flows are what’s required to service and ultimately de-lever the high debt loads being stacked onto balance sheets today.
What it means for your portfolios
As we evaluate the investment implications of the AI Supercycle, the most critical takeaway for clients and advisors is that the AI infrastructure buildout is a dynamic, rapidly evolving credit story rather than a static macro trade. Fixed income investors cannot afford to paint the AI ecosystem with a single broad brush. Within the hyperscalers alone, there is a vast credit difference between those with fortress balance sheets such as Microsoft, Alphabet, Amazon and Meta, and a riskier borrower like Oracle.
For our portfolios, this divergence is precisely where active fixed income management shows its value. Rather than chasing the AI theme or buying new supply just because index weights are rising, we focus on balance sheets, leverage and free cash flow, and we own an issuer only where the spread pays us for the risk. We seek to capitalize on opportunities to lock in yields on dominant, high-margin franchises, while avoiding more levered borrowers whose cash flows are highly dependent on unproven monetization timelines.
As this historic capital cycle continues to unfold, careful credit selection will be a primary tool for capturing the upside of the AI transformation while managing the evolving risks.
This commentary is provided for informational and educational purposes only. As such, the information contained herein is not intended and should not be construed as individualized advice or recommendation of any kind.
The opinions and forward-looking statements expressed herein are not guarantees of any future performance and actual results or developments may differ materially from those projected. The outlook for hyperscaler free cash flow recovery depends on successful monetization of AI investments; weaker demand, pricing pressure or continued high capital spending could delay or prevent that recovery. The information provided herein is believed to be reliable, but we do not guarantee accuracy, timeliness, or completeness. It is provided “as is” without any express or implied warranties.
All investing involves risk, including possible loss of principal. Fixed-income investments are subject to interest rate, credit, liquidity and call risks, and bond prices generally decline when interest rates rise. An investment-grade rating does not eliminate these risks. Active management and credit selection do not guarantee a profit or eliminate losses, and even financially strong issuers may experience credit deterioration or bond price declines.
Mariner is the marketing name for the financial services businesses of Mariner Wealth Advisors, LLC and its subsidiaries. Investment advisory services are provided through the brands Mariner Wealth, Mariner Independent, Mariner Institutional, Mariner Ultra, and Mariner Workplace, each of which is a business name of the registered investment advisory entities of Mariner. For additional information about each of the registered investment advisory entities of Mariner, including fees and services, please contact Mariner or refer to each entity’s Form ADV Part 2A, which is available on the Investment Adviser Public Disclosure website. Registration of an investment adviser does not imply a certain level of skill or training.