Time for private credit to stop grading its own homework

By Terence McMenamin, co-founder and CEO of Techdollar

There is a new vertical of credit being unlocked, and it is the largest one nobody has built for. A decade of companies staying private has left employees, founders, and early investors holding trillions of dollars in equity in the most important companies on earth, with no clean way to borrow against it. The frontier-yield thesis is the lending side of that same gap: real-world, largely non-correlated credit in the double digits, sourced from the actual economy rather than from anything reflexive, drawing on a financeable market north of $30 trillion. The thesis is right. I want to sharpen the part it carries with it.

Moving capital was never the hard problem. Underwriting it was. Every credit wrapper in this lineage went through one detonation before it matured, and the detonation here will not be a liquidity event in the usual sense. It will be a marking event. The only question that decides who is still standing afterward is whether the collateral behind the yield can be priced when it counts, or only appraised when the market is calm.

Underwriting was the hard part

Distribution, settlement, speed: those got solved. A 24-hour quote in place of a four-week process is real progress, and it is also the easy half. The hard half of private credit was never moving the money. It was knowing what the loan is worth on a bad day.

Private credit marks by appraisal, not by tape. That suppresses volatility in normal years and only discloses the realized loss curve in the tail. The Reserve Fund broke a buck because Lehman paper had been carried at par. Greensill ran short-tenor paper that looked conservative until it didn't. In the first quarter of last year, Blue Owl took close to 22 percent redemption requests on a $36 billion flagship because the market could not price its software book, and Jamie Dimon used his shareholder letter to say it plainly: private credit does not tend to carry rigorous valuation marks on its loans.

Most of the new credit being underwritten inherits exactly this. Direct-lending books, supply-chain paper, the long tail of private assets. Appraisal-marked, model-priced, and opaque on the one day the mark has to hold.

The exception has a tape

One frontier-yield collateral does not have the marking problem: late-stage private tech equity.

The companies that anchor private markets today are not pre-revenue bets on unproven markets. SpaceX, OpenAI, Anthropic, Stripe and their peers are production systems with hard assets, enforceable IP, and multi-year demand. And for the first time, their equity trades. Jefferies put global secondary volume at roughly $240 billion in 2025, a record. There are about 1,500 pre-IPO unicorns now, up from 39 in 2013, and the median time to IPO has stretched past eleven years. Caplight, Hiive, Setter, Secondlane and a deep OTC network print observable prices on this paper continuously. Caplight is, in effect, CoStar for private markets.

That is the entire difference between marked-to-model and marked-to-tape. Frontier tech equity is one of the only private collateral classes you can price on a Tuesday afternoon at a level someone else will transact at. The industry still lumps it in with mid-market software under one "private credit" label and prices both like venture, pricing in a high chance of default and hoping the upside covers it. The tape says otherwise. A loan against a SpaceX or an Anthropic is not a venture bet with a coupon. It is a secured loan against the infrastructure of the next economic cycle.

Underwrite to the tape

This is the credit vertical we are building. Holders of equity in the most important private companies pledge that equity and borrow against it, underwritten near 40 percent loan-to-value against live secondary marks rather than a quarterly appraisal. Ownership and the security interest are verified by RSM, a top-ten CPA firm, before a dollar moves. When a loan has to be unwound, it routes through the same venues that produced the mark in the first place.

Continuous pricing instead of periodic valuation. A real liquidation path instead of a model. Repayment at tenor, not pinned to an IPO or an exit that may never arrive on schedule. The return that reaches lenders comes from the loan book itself: interest on secured credit, priced off a collateral value you can observe.

And the cost of capital reflects that collateral. Borrowers access funds at roughly 12 percent all-in where the closest alternatives charge 19 percent and up, funded in 24 to 48 hours against an incumbent process that takes a month and declines anyone holding less than $20 million. When you can price the collateral, you can price the loan. Everyone else is pricing blind and charging for the privilege.

Built onshore, on purpose

We are building this in the United States, deliberately, as enforceable US credit. Every loan is a real legal instrument: a perfected security interest in verified collateral, ownership attested before funding, a defined path to enforce. That is a choice. The easier version of this business lives offshore, in lighter-touch jurisdictions, one step removed from the assets and the courts that stand behind them. For a credit business whose entire edge is legal certainty, that is the wrong trade.

It is also the right side of where the economy is going. The collateral we lend against is the American industrial base in private form: launch infrastructure, frontier AI, domestic compute and silicon, robotics, defense. Capital is rotating back toward atoms, and the companies building those atoms are overwhelmingly being built here. Credit that backs them should be built here too, under US law, in step with the country's own push to fund the next industrial cycle rather than routed around it.

The honest risks

I would rather name them than route around them. The secondary can thin. Marks can go stale in a true freeze, which is precisely when you need them most. Concentration is real, because a book of correlated frontier names moves together. And these assets are volatile even at the very top of the market.

The discipline is the answer, not the denial. A mark you will defend in a downturn is a multi-venue, volume-weighted reference with explicit staleness handling, not a single optimistic print. Concentration gets capped before it is a problem, not after. Loan-to-value stays conservative for one reason: 40 percent exists to survive a 50 percent drawdown with collateral to spare.

Then there is the tail question everyone asks. What happens if these assets fall by half or more? Ask the larger version of it. If the SpaceX's and the Anthropic's of the world go to zero, a private credit book is the least of anyone's concerns; that is an existential problem for the economy, not a credit problem. Short of that, in a genuine drawdown credit becomes the superior trade. No one wants to abandon the thesis they are building, least of all the people building these companies. But everyone still needs capital. That is exactly the moment a secured, properly marked loan becomes the most attractive instrument in the stack.

The narrow bet

Frontier yield is the right call, and the bet inside it is narrower than the category. The durable version is not private credit on a rail. It is private credit you can mark on the rail. The collateral that already trades is where the yield survives the cycle that every wrapper eventually meets.

The old wealth strategy was buy, borrow, die. The frontier version is build, borrow, thrive. You earned the equity. You should not have to sell it to use it, and the people lending against it should be able to see exactly what it is worth.

Solve distribution and you have a feature. Solve the mark and you have a credit business that is still standing after the detonation.

Terence McMenamin is co-founder and CEO of Techdollar, the credit layer for late-stage private equity. He previously architected institutional structured products at Dinero Labs, acquired by Plume Network in 2025, with backing from Galaxy and Nomura.