If you have a 401(k) and haven't touched the settings, there is a good chance you are now lending money to fund AI data centers. You didn't choose to. No one asked. The mechanics of how retirement funds work made it happen on their own.
Here's what to know.
How the money gets in
Target $TGT-date funds held about $4.8 trillion in assets at the end of 2025, with the total market tracked by Sway Research reaching $5.2 trillion when custom strategies are included. The Pension Protection Act of 2006 facilitated their rise, making target-date funds an approved default investment for employees automatically enrolled in 401(k) plans. By 2020, 84% of plan participants were offered target-date funds, according to a study cited by the U.S. Government Accountability Office.
These funds follow a "glide path." They start with a heavy allocation to stocks and shift over time toward a mix weighted more toward bonds. These funds are designed so you can put your money in and let the managers handle the rest. The bond portion of a target-date fund is where the AI connection enters. Each Vanguard Target Retirement Fund, for instance, invests in Vanguard's broadest index funds, giving investors access to thousands of U.S. and international bonds.
The dominant benchmark underlying those bond index funds is the Bloomberg U.S. Aggregate Bond Index, known as "the Agg." It is a broad, market-capitalization-weighted bond market index that tracks government debt, investment-grade corporate bonds, mortgage-backed securities, and asset-backed securities. Funds that attempt to replicate its performance include the iShares Core U.S. Aggregate Bond ETF (AGG), the Vanguard Total Bond Market Index Fund, and the Fidelity U.S. Bond Index Fund. The critical word is "market-capitalization-weighted."
Bonds in the index are weighted based on how large each type is in the market. To make the cut for the Agg, a bond needs to clear three bars: it must carry an investment-grade rating, have at least $100 million worth outstanding, and still have a year or more left before it matures.
The composition of the returns universe is rebalanced at each month-end; a forward-looking projected universe changes daily to reflect issues entering and leaving the index, and on the last business day of each month, that projected universe becomes the returns universe for the following month. That rebalancing is the hinge. When a company issues a large volume of investment-grade bonds, those bonds enter the index at the next rebalance. The index's weight in that company's debt rises. Every fund tracking the index must then buy more of those bonds to match. No human at your 401(k) provider decides this. The index rules dictate it.
Why the index is changing fast
This is where the AI buildout enters. Tech companies are now taking their place among the largest borrowers in the American credit system, issuing bonds at a scale that is reshaping the corporate bond market. Hyperscaler gross bond issuance topped $100 billion in 2025, according to the Bank for International Settlements, with most issuance long-term, locking in funding for multi-year build-outs.
Companies including Alphabet $GOOGL, Amazon $AMZN, Meta $META, Microsoft $MSFT, and Oracle $ORCL collectively issued $100 billion of bonds in 2025, more than double the amount they raised in the prior year, according to Bank of America $BAC. The result is a direct shift in index composition. According to research from Breckinridge Capital Advisors, Meta, Alphabet, Amazon, and Oracle's collective weighting in the Bloomberg U.S. Corporate Investment Grade Index nearly doubled over the year ending April 1, 2026, from 2.2% to 4.1%. AI-linked issuance was about 5% of total investment-grade corporate bond issuance in 2025, about three times the prior decade's average annual tech issuance.
Apollo chief economist Torsten Sløk has projected that if hyperscalers finance even 20% of their planned capital expenditures through investment-grade bonds over the next four years, the rest of the index growing at 5% per year, the rankings of the largest bond issuers would reshuffle. Amazon would vault into the top three. Meta, Microsoft, Oracle, and Google would all push into the top 10.
"When you start issuing $25 or $50 billion dollars of debt at a clip, it's very easy to put yourself into that top 10 largest borrowers," Jason Greenblath, a senior portfolio manager at American Century Investments, told Quartz in a recent interview. "So the composition of the index will change. It is changing."
Because the Agg is market-cap-weighted, every dollar of new tech debt that meets the eligibility criteria increases that issuer's weight in the index. Most bond indexes are weighted by market capitalization, which results in what analysts have called the "bums" problem: Issuers with a lot of outstanding debt constitute a larger part of the index than those with less debt. Funds tracking these indexes have no choice but to follow.
What this means for your 401(k)
The chain works like this: Tech companies issue bonds to finance AI infrastructure. Those bonds are investment-grade and meet index inclusion rules. Routine monthly rebalancing adds eligible new issues and removes matured or noncompliant bonds. Bond index funds tracking the Agg buy the new issues proportionally. Target-date funds hold those bond index funds. Your 401(k) holds the target-date fund.
At no point in this chain did you make a decision about AI-related debt. About 61% of target-date fund assets at the end of 2024 were in passive strategies, according to Sway Research, meaning they track indexes mechanically. Vanguard alone managed $1.79 trillion in target-date assets at the end of 2025, growing by $310 billion that year.
"Some of this paper ends up in broad-based corporate bond funds that sit in target-date retirement products," independent analyst Dave Friedman told Quartz. "The retail holder has no idea why they own it, no framework for evaluating it, and no ability to exit it independently of the fund."
That is not a scandal in itself. Diversified bond funds have always held things individual investors don't track. But the speed and scale of AI-related issuance mean the composition of what you own is changing faster than the information available to you.
The disclosure gap
The Department of Labor developed guidance on target-date funds over a decade ago, and that guidance has not been updated. The DOL created guidance in 2010 for participants and in 2013 for plan sponsors. The guidance does not include recent developments such as the increase of target-date funds structured as collective investment trusts, which are bank-administered pooled funds with different disclosure requirements than mutual funds.
In a March 2024 report, the GAO recommended that the DOL update both documents. The DOL disagreed with both recommendations. The underlying issue is structural. An investor looking at a target-date fund's fact sheet sees broad labels such as "investment grade corporate." They do not see how the weight of tech-sector debt within that category has shifted, or that a growing share of new issuance in those categories is financing a specific asset class with untested residual economics, as Friedman put it.
Whether any of this is a problem depends on your view of the underlying credit. As Quartz's reporting on the broader AI debt story makes clear, most hyperscaler bonds are backed by strong cash flows and carry investment-grade ratings for good reason. The risk profile of GPU-collateralized neocloud debt is different. Both types of issuance can flow into the same broad indexes and, from there, into the same retirement accounts.
The point is not that your 401(k) is in danger. It is that the contents of your 401(k) are changing, and the disclosure regime has not kept pace with the speed of that change. Knowing that is the first step toward evaluating whether it matters to you.