New Listings: An AI Bond Fund Will Lend to Publicly Traded Companies, and Its List Includes OpenAI and Anthropic

A new bond fund says it invests in debt issued by “publicly-traded companies“. Two paragraphs later, it names OpenAI and Anthropic as eligible issuers.

Defiance AI Infrastructure Bond ETF (AIBN, NYSE Arca)

I recently described a fund that reaches OpenAI and Anthropic through swaps on perpetual futures, because there are no shares to buy. Defiance has found the other door. If you cannot own the equity, you can at least lend the company money.

The MarketVector US Hyperscalers and AI Bonds Index sorts its issuers into three groups. The four “Hyperscalers" are Amazon, Microsoft, Alphabet and Meta. The “Other Compute Infrastructure and Large Language Model Companies" are IBM, CoreWeave, Oracle, SpaceX, OpenAI and Anthropic. Everyone else comes from a semiconductor index, which requires new entrants to earn at least half their revenue from chips.

The weights are fixed with some precision. The hyperscalers receive 48% of the index between them, 12% each. The LLM and compute group may have up to 20%. Semiconductors receive at least 32%. Each bond needs at least $1 billion outstanding, and both investment-grade and high-yield bonds are welcome.

The prospectus is candid about what the investor gives up. The fund “does not participate in an issuer's equity appreciation". So, the most excitable corner of the stock market has been turned into a fixed coupon, with nearly half of it lent to four companies that already have more cash than most governments.

The eligibility list is aspirational for now. Private placements and Rule 144A bonds, where a private AI laboratory’s debt would most likely sit, are eligible but not currently included. The expense ratio is 0.39%.

Hartford High Yield ETF (HINC, Nasdaq)

Not every new ETF is new. This one is the Hartford High Yield Fund, a mutual fund, converted into an ETF with the same objective, the same Wellington sub-adviser and the same strategy.

It inherits a ten-year performance record from its predecessor, with one footnote: for most of that decade the fund “pursued a modified investment strategy", and its current portfolio manager arrived only when the strategy changed. At least 80% goes into high-yield bonds, up to 20% into bank loans, and the net expense ratio is 0.50%.