🧠 Private Credit is Cooked*

The SaaSpocalypse Hits Private Credit

The world's largest asset manager just paused redemptions on its flagship private credit fund.

That’s the kind of sentence that gives me financial crisis heebie jeebies. 

BlackRock's $26 billion HPS Corporate Lending Fund received $1.2 billion in redemption requests. It couldn't honor roughly $580 million of them. The fund was gated. Investors who thought they had quarterly liquidity discovered they didn't.

That same week, 

Ruh-roh.

Redemption requests across the industry are running at roughly 8% this quarter. In 2024, the average was 1.3%.

The $3 trillion private credit industry is getting its first real stress test. 

And the reason isn't what most people think.

*Actually it’s not yet cooked, it’s going through a major re-pricing in its most concentrated sector exposure and we don’t know yet if that’s contagion. Maybe, partially baked is more apt.

For a quick recap on private credit:

In that same piece I called private credit funds "efficient risk transfer machines."

I was right about the machine. I was wrong about what it was carrying.

And what it was carrying was a lot of SaaS exposure.

Software is private credit's single largest sector exposure. 

Software and technology companies account for roughly 25% of the private credit market. UBS puts the AI-disruption-exposed share even higher, at 25-35%.

That concentration exists because the SaaS thesis was, for a decade, the perfect lending story. 

Annual Recurring Revenue (ARR) was the perfect investment. 

From 2015 to 2025, more than 1,900 software companies were acquired by private equity in deals worth over $440 billion. Private credit loved lending against those characteristics. Software became the engine of the entire unitranche loan market.

AI broke the thesis.

Every one of those assumptions — sticky revenue, high margins, switching costs — is being stress-tested simultaneously. 

The idea that these loans were “safe” has shattered and investors want out. 

And just to add to the misery, the loans were underwritten when SaaS companies were being acquired at 24x multiples. By 2025, that had compressed to 18x. Today the median sits around 16x. 

Many of those 2021-2022 vintage deals were underwritten at valuations that have since been cut in half. The loans are still being repaid, but they’re also being repriced as SaaS stocks tumble.

The selloff has erased more than $265 billion in market cap from the alternative asset managers. 

And the real exposure may be worse than reported. Bloomberg found that at least 250 loans to software firms worth more than $9 billion were categorized as other industries by BDCs. The sector concentration is deeper than the headline numbers suggest.

In my earlier piece I wrote:

The investors who "signed up for it" thought they were buying stable SaaS cash flows. 

They were buying business models being eaten alive by generative AI.

Every cycle has the same pattern. Wall Street creates an asset class. Institutions get in early. Returns look great. Then they open the door to retail right as the cycle matures.

BCRED, HLEND, Blue Owl's retail vehicles were designed for "wealthy individuals."  Also known as retail. And on the surface they look like savings probdocuts with better yield. 

They behave like illiquid credit portfolios with concentrated sector exposure.

The conventional wisdom is that private credit is such a good asset class we should be giving it to everyone in retail. We should democratize it. Why should the institutions have these stable, long-term, high returns? 

When the CEO of Robinhood announced Robinhood is launching Robinhood Ventures Fund I, a closed-end fund giving retail investors exposure to private companies like Databricks and Ramp. The CEO said:

It dropped 11% on day one. 

I'm not against retail having access to private markets. I'm for retail understanding what they're buying. And right now, the gap between the pitch and the product is uncomfortably wide.

Last week, JPMorgan marked down the value of software loans held as collateral by private credit funds — and restricted lending against them. The bank went through its financing portfolio "name by name, sector by sector." Jamie Dimon told investors JPMorgan was being "more prudent in lending against software assets." 

When the bank that provides back-leverage to the lenders starts marking down their collateral, the tightening cascades. Less leverage means less lending capacity means more pressure on already stressed portfolios.

So what you actually have is a scenario that looks like this:

Just to give you a sense of scale here. Private credit is $2-3 trillion. The global bond market is $130 trillion. Banking assets top $180 trillion. Even the largest private credit funds cap out at $20-80 billion. 

The thing to watch isn't a single fund blowing up. It's confidence. 

If software loan losses make capital allocators gun-shy about private credit broadly, the repricing spreads to healthy loans too. That contagion path — from SaaS to broader credit — is where this gets interesting.

Probably not.

Unlike in 2008, it’s much harder to have a run on the private credit funds like a run on a bank. The pausing of redemptions creates a dynamic where even if there is a “run on the funds,” they can smooth that out over time.

Maybe the Robinhood launch was a case of bad timing. The asset class is going through a macro re-pricing of its most concentrated sector.

Private credit will survive. It'll reprice, create new funds, and underwrite a new reality where AI exists. But right now, the repricing means less lending capacity — right when AI labs and hyperscalers are thirsty for capital.

That's a monumental window for banks.

The regulations that pushed banks out of risky corporate lending post-financial crisis — and created a structural opportunity for private credit — are softening. 

Banks are getting their lane back. 

Will they be willing to finance the hyperscalers' giant data center build-outs? Do they have that muscle? Time will tell.

The irony is rich. Post-2008 regulations pushed risky lending out of banks and into private credit markets. Private credit gorged on it, particularly in software. AI cracked the thesis. And now the regulations that created the opportunity are being rolled back — just as the opportunity sours.

But the borrowers in those 2021-2022 vintage funds? 

The retail investors in semi-liquid vehicles they didn't fully understand?

They're the ones who'll feel it.

ST.

1. Agent Card - Prepaid virtual cards for AI Agents

Agent card is an API service to quickly give AI agents cards via the command line. It has 8 simple commands to create cards, check balances, login and out etc. Cards are issued instantly, and work everywhere major card brands are accepted. 

🧠 Why is this different to regular virtual cards? Simply put, its built for agents first. Every other virtual card issuer supports humans first, agents second. Check out their “are you an agent” page. Sometimes simplicity is everything, and for the customer of tomorrow, agents, thats absolutely true. 

2. Burst - HSA/FSA cashback API

Burst helps consumers use their health insurance to make wellness purchases, without having to check whats eligible or fill in forms. Merchants install an API, that runs post purchase. Then the customer connects their HSA/FSA provider, Burst will automatically make the claim on their behalf. They also have a browser extension to quickly see what’s eligible regardless of store. 

🧠 I’ve seen plenty of cashback cards, but I hadnt seen the merchant integration side. This feels like a way to get more sales for an API integration. Not trivial for merchants, but probably ROI positive.

3. Payr - Bilt for the UK (kinda)

Payr helps tennants pay rent via their existing credit card, while landlords receive traditional bank transfers. Users enter their tenancy information to the app, add their preferred card, and can pay rent and collect reward points or cashback. 

🧠 How do they make money? Payr charges a 1.99% payment execution fee on each rent payment. Remember, UK domestic credit card fees are capped at 0.3%, so there is potentially some spread here. Interestingly they’re targeting business renters as well as consumers. And “payr points” are coming soon. Think about international students trying to pay rent via SWIFT. That alone is an interesting wedge.

4. Cyclops - The Stablecoin Platform for Payments companies

Cyclops helps payments companies add stablecoin use cases like payouts, collections, settlement and conversion. The platform is optimized for payments businesses to integrate with their existing reconciliation and compliance workflows.

🧠 Cylops is a spin out from the Shift4 Fintech team. Payments companies entering stablecoins face a patchwork of vendors, and a lot of self-build like reconciliation or compliance workflows. 

1. Mastercard acquires stablecoin infrastructure BVNK.

Mastercard will acquire BVNK for $1.8bn to “connect on-chain and off chain finance.” BVNK enables stablecoin payments for companies like Worldpay, Flywire and Dlocal, in over 130 markets, with local licences and compliance capabilities. The press release cites increasing demand from institutions, regulatory clarity and the real-world data showing stablecoin payments are powering at least $350bn of volume. The Mastercard CPO expects most institutions will come onchain.  

🧠 Mastercard gets a capability, BVNK gets distribution. Plenty of banks already have a commercial agreement with Mastercard. If they can switch on the sales engine, there’s plenty of potential revenue upside. 

🧠 Every Bank is currently shopping for vendors and partners to get into this space. For orchestration it three names consistently rise to the top: Bridge, BVNK, ZeroHash.

🧠 Bridge had a larger parent in Stripe who could help with distribution. ZeroHash had many deep tradfi partners and investors. BVNK was punching above its weight class and hanging in. Now it can get way more distribution.

🧠 Every bank will go onchain. It’s a question of when. The Mastercard CPO’s statement that they expect that to happen could be viewed skeptically as, they’re trying to sell their new product. But I don’t think Mastercard makes this acquisition unless there’s demand.

🧠 Cards are winning in stablecoins. Visa’s fastest growing revenue line is stablecoin-linked cards. Far from displacing the card networks, stablecoins are another funding source. 

🧠 The most disruptive thing nobody is paying attention to is stablecoins for card network settlement. Visa doesn’t move money banks do. Cross-border this is slow and expensive with the correspondent banking network. Both networks are making stablecoin settlement an option. (BVNK powers Visa’s direct settlement btw, wonder what happens there)

🧠 BVNK was an early high-conviction bet on stablecoins that paid off. They did the hard yards getting licences and winning early enterprise deals, now at a $30bn annual run rate of real payments flows. Huge congrats, Jesse, Chris, Emma & the team.

1. Compute will be a top 10 commodity in 5 years.

The Founder of Chicago-based prop trading firm DRW says compute will be world's top commodity in 10 years. People will spend more on GPUs than an oil, which means that there should be a futures financial market for GPUs. 

🧠 There’s a tiny handful of people talking about this. And yet, it feels inevitable. Worth a read.

Remember, if you're enjoying this content, please do tell all your fintech friends to check it out and hit the subscribe button :)

Want more? I also run the Tokenized podcast and newsletter.

(1) All content and views expressed here are the authors' personal opinions and do not reflect the views of any of their employers or employees.

(2) All companies or assets mentioned by the author in which the author has a personal and/or financial interest are denoted with a *. None of the above constitutes investment advice, and you should seek independent advice before making any investment decisions.

(3) Any companies mentioned are top of mind and used for illustrative purposes only.

(4) A team of researchers has not rigorously fact-checked this. Please don't take it as gospel—strong opinions weakly held

(5) Citations may be missing, and I’ve done my best to cite, but I will always aim to update and correct the live version where possible. If I cited you and got the referencing wrong, please reach out