There is a peculiar job in private credit: checking the person you hired to do the checking. A fund appoints a loan administrator to calculate interest, record payments, and prepare reports. Then its own staff maintain a second ledger, often in Excel, to make sure the first one is right. The industry calls this shadow booking. It is an elegant name for paying twice to acquire confidence once.
- Hypercore manages the operational life of private loans.
- Its newer service combines AI execution with human oversight.
- The useful buying test: bring a complicated deal to the demo.
Hypercore has chosen this small, stubborn absurdity as its opening. The company supplies loan management software to private credit funds and non-bank lenders. In 2026, it added a managed administration service, its AI Admin Agent, that performs servicing work inside the same system. The ambition is straightforward: fewer parallel records, fewer handoffs, and calculations a lender can inspect without rebuilding them.
01 / The loan that refuses to fit
A private loan can be a remarkably individual object. Interest may arrive partly in cash and partly as payment-in-kind, added to the balance. Capital may be drawn in stages. Several lenders may share a facility, with different allocations and a payment waterfall governing who receives what. A tidy dashboard has little value if the underlying system cannot describe the agreement.
Hypercore’s platform brings deal terms, schedules, borrower information, funding sources, and reporting together. Its customers span direct lending, venture debt, and commercial real estate. The point is to give an operations team somewhere to represent the loan it actually negotiated, then automate recurring calculations and notices. Software built around a standard borrower can become awkward when almost every agreement contains an exception.
02 / First, build a place for the numbers
The founding team came to lending through software development. While building fintech products, a client asked them to support a lending product. They expected suitable infrastructure to exist; their account of the discovery is that bank-oriented systems did not fit private lenders. Hypercore grew from that gap, combining software experience with the daily mechanics of debt.
Founded in 2020, it entered Y Combinator’s Summer 2021 batch. Daniel Liechtenstein, Tomer Moshe, David Yahalomi, and Eitan Frailich had also built Articode, a development business. The early YC description focused on smaller lenders, workflow automation, and portfolio analytics. The subsequent emphasis on private credit gives the product a more exacting customer: a fund whose tailored agreements must become dependable operational records.

03 / Let the awkward deal into the demo
At Pinegrove, director of operations Christopher Kong needed independent loan infrastructure after the business separated from its former parent. Existing systems and outside providers left too much manual work as the portfolio grew. His buying decision turned on a practical courtesy: Hypercore let him test example deals and amortization schedules during the sales process.
“The cost of keeping things as-is was greater than the cost of making a transition to a new system.”
Christopher Kong · Pinegrove · published customer account
Hypercore’s case study reports implementation within weeks and lower servicing costs. The transferable lesson is the test drive. Ask a prospective vendor to model a difficult agreement and explain its resulting schedule. A sales presentation can glide over an exception; a calculation has to land somewhere.
04 / Buy before the spreadsheet grows teeth
FH Structured Solutions took a different route. When Frazier launched the healthcare-focused credit strategy in January 2025, the team adopted Hypercore from the beginning. Its published case describes mixed cash and PIK interest, variable amortization, and warrant components. Starting early avoided migration and backfilling; automation reduced the need for additional operations staff.
Greymax, a UK property bridging lender, arrived after spending a year unsuccessfully trying another provider. Its problem involved bespoke loans, staged drawdowns, and investor reporting. Hypercore’s account says the team implemented its platform in under six months. These are vendor-published experiences, rather than promises that every installation will follow the same timetable. They still expose something useful: implementation deserves as much scrutiny as the feature list.
05 / Better software left a job unfinished
Hypercore’s March 2026 explanation of its change in direction is unusually revealing. Customers had better software, yet still waited for requests, reconciled manually, and kept shadow books. The platform had improved the work without removing the obligation to perform it. That residual workload became the case for delivering administration as a service.
The AI Admin Agent became generally available in May 2026. Hypercore describes agents carrying out onboarding, servicing, reconciliation, and reporting on the platform, with professionals validating outputs and handling exceptions. Clients can appoint the service for a deal, selected facilities, or a portfolio, and can choose software, managed service, or both. For a buyer, the distinction is who owns the daily execution.
06 / Capital for the unglamorous part
In February 2026, Hypercore announced a $13.5 million Series A led by Insight Partners, with Atinc and Y Combinator continuing their support. The company reported more than $20 billion across over 10,000 loans on its platform, and 3.5-fold growth in contracted annual recurring revenue during 2025. Those figures describe customer portfolio coverage and contracted revenue growth; they do not make Hypercore a $20 billion lender.
Led by Insight Partners
The June integration with Arcesium extends the same logic beyond the loan ledger. Hypercore supplies loan-level operations; Arcesium supplies fund operations and data infrastructure. Their connection is designed to carry loan events into accounting, net asset value calculations, reconciliation, and investor reporting. Removing a spreadsheet in one department means little if another department has to recreate it.
07 / Keep the review; lose the relay race
A September 2026 release makes the proposition tangible. Hypercore’s statement workflow combines configuration, review, sending, and tracking. Documents must be reviewed before delivery, with permissions separable across those stages. Errors stay visible for correction. The company’s claimed reduction from days to about an hour is a product claim, but the design lesson stands: automate coordination while keeping the approval legible.
Hypercore fits where bespoke lending creates repeated operational work. Buyers should test their loan structures, migration requirements, permission rules, and exception handling before entrusting execution to it. A fund whose existing processes already work well has a different calculation to make. The question is how much time its people spend maintaining confidence in the numbers - and whether a second spreadsheet is still the price of admission.