Hyperscaler bond issuance pushes tech past banks in IG bond indexes

Alphabet, Amazon, Meta, Microsoft and Oracle borrowed more than $132 billion in the first half of 2026 to fund AI infrastructure, enough that technology now outweighs banks in the indexes behind plain-vanilla corporate bond ETFs. The bond market is financing the buildout at a scale that has begun to redefine what a broadly diversified investment-grade portfolio holds.

Technology weight in pure corporate bond ETFs rose about 300 basis points, according to ETF Trends, enough that the sector passed banks in major investment-grade bond indexes, and for an investor who bought a broad corporate bond fund for ballast, the underlying risks changed while the fund's name stayed the same. The bonds financing data centers, chips and related AI infrastructure now sit where financial-sector debt used to lead.

In a market-value-weighted bond index, an issuer's weight rises when its outstanding debt grows, and the five companies issued at a pace that forced the benchmark to reallocate. No investor had to buy a technology credit fund and no portfolio manager had to take a view on AI infrastructure; the index did the work.

A bond index ranks issuers by their outstanding debt, and when a company sells a new bond, that debt joins the index on its settlement date. If the company is already among the largest borrowers, the new bond adds to its weight rather than displacing it, producing a snowball effect where large issuers get larger by borrowing more, not by becoming more creditworthy.

For the five issuers in question, the aggregate of more than $132 billion across Alphabet, Amazon, Meta, Microsoft and Oracle meant that the largest corporate bond benchmarks had to absorb a huge slab of new paper from a single sector, which is the opposite of diversification but exactly what a passive index is designed to do: mirror the market as it is.

Bond ETFs are often treated as the stable side of a client portfolio, but an advisor using a broad investment-grade bond ETF for diversification now holds more credit risk tied to the same AI capital-spending cycle that has lifted technology equity indexes, so the equity and fixed-income sleeves may now share an industrial exposure they did not share by choice.

The concentration is not evenly distributed across equity indexes: T. Rowe Price counts AI as nearly 60% of the Russell 1000 Growth Index, with hyperscalers and AI infrastructure concentrated in large-cap growth and the lightest exposure in U.S. small-caps and EAFE value, which gives equity investors at least a visible choice—move down the market-cap spectrum or shift geography. The bond index offers no such obvious alternative because the same five issuers dominate the investment-grade corporate market.

European flow data from the week to September 25 shows what an investor-driven tech trade looks like: Information Technology led sector performance at 4.57% and drew €275.7 million, while financials shed €1.03 billion, money following returns as it often does in equity sector funds. The bond index shift needed no such flow; the weight rose because the bonds were issued and added to the benchmark, independent of any investor selling banks or buying tech credit.

European financial-sector ETFs shed €1.03 billion in the same week, but in the bond index banks did not lose their top spot because investors sold financial credit; they lost it because technology issued faster than banks did. A sector rotation in an ETF implies a buyer and a seller, while an index sector crossover implies only an underwriting calendar and a benchmark rule.

The difference is between a demand story and a supply story: sentiment expressed through trades versus arithmetic expressed through issuance schedules. Advisors who treat the two as the same kind of exposure change may be missing the difference, because a bond ETF shareholder cannot vote on whether Oracle's leverage is appropriate or whether Meta's data-center spending will earn its cost of capital; the fund holds the bonds at market weight until the index says otherwise.

Passive fixed-income investors now hold AI credit risk without choosing it; an investor who wanted to underweight technology credit or avoid the hyperscaler capital-spending cycle has no way to do so inside a broad bond index fund except by leaving the index entirely. Equity managers can debate trimming growth exposure and active bond managers can avoid specific issuers, but the benchmark-tracking fund is fixed by the rulebook.

When issuance is concentrated among five borrowers, the index becomes more concentrated even if the number of issuers stays the same, and a corporate bond index can cross a sector threshold simply because a handful of companies tapped the bond market at scale rather than because the sector's fundamentals improved. Banks slipped behind tech because tech issuance grew faster than bank issuance, while bank balance sheets held steady.

The five companies are not marginal borrowers; their combined first-half issuance is the financing side of a capital-spending plan that spans data centers, power, chips and networking. At that scale, the resulting index weight is concentrated by design: each new deal adds to the stock of debt that the index must track, and the only offsets are slower issuance, maturing bonds, or enough debt from other sectors to dilute the share.

If the same issuers return to the bond market at a similar pace in the second half of 2026, the technology weight in pure corporate bond ETFs will likely keep climbing, because each new bond deal adds to the outstanding debt that defines the sector's share. Those offsets are not levers available to the investor and not decisions a bond ETF shareholder can make.

The bond sleeve's diversification benefit may be narrower than the fund's label suggests: if the equity sleeve is heavily weighted to megacap tech through a growth index and the bond sleeve is now also overweight technology credit, the client's total portfolio is more exposed to one industrial story than either allocation alone would suggest. That is arithmetic across two sleeves.

The first half already changed the benchmarks; technology now outranks banks, and the buyer never had to sign off. The composition fact carries a dollar figure: $132 billion.