Four instruments fund the same growth. Only one of them gets more expensive when you do well.

What is the actual choice, and why is it hard to make?

Four instruments funding one need, quoted in four incompatible units. Revenue-based financing quotes a multiple. Venture debt quotes a rate and a warrant. A recurring-revenue term facility quotes a spread and a covenant package. Growth equity quotes a percentage of the company, permanently.

Nothing about that list is exotic, and yet the comparison is routinely made badly, for a structural reason: the four are not offered to the same company at the same time by the same people. Each arrives through its own channel, with its own vocabulary, at its own moment, and the founder ends up choosing between the one on the table and nothing. The discipline that fixes this is to hold the capital need constant, name the four candidates before talking to any of them, and require each to be expressed in the same units as the others.

One vocabulary trap deserves naming first, because it costs real money. Fixed-share revenue-based financing and recurring-revenue term facilities are both sold as lending against recurring revenue, and they are not the same instrument. One takes a percentage of revenue until a fixed multiple is repaid and carries no covenant package worth the name. The other is conventional debt sized off annual recurring revenue, with a spread, a maturity and covenants that convert to earnings tests as the company scales. Sorting a borrower between them is the first advisory act, not the last.

What does revenue-based financing require, and what does it cost?

Recurring revenue, and very little else. There is no sponsor requirement, no covenant package in the conventional sense, and no fixed maturity. The provider advances an amount and takes a share of monthly revenue, often 5% to 15%, until a fixed cap has been repaid, with caps commonly running 1.5x to 3x the funding amount (Re:cap, revenue-based financing guide, published 28 August 2026, accessed 6 September 2026).

The price behaves unlike anything else on this list, and the direction is the whole point. Because the cap is fixed, the dollars repaid do not change with speed; only the time does. Faster revenue growth means faster payments, earlier repayment and a higher implied return to the provider (Re:cap, same source). A company that beats its plan pays a higher annualized rate precisely because it beat its plan. Every other instrument in this comparison rewards the same outcome.

That is not a case against the instrument. It is a case for pricing it properly before signing, which is a separate piece of work and is set out in full in the companion article on what the repayment cap actually costs. The short version for the purposes of this comparison: a multiple is not a rate, and a borrower who has not converted one into the other cannot compare this option to the other three at all.

Why is venture debt closed to most of the companies that want it?

Because the underwriting is the syndicate, not the company. Venture debt as conventionally practised sits alongside institutional venture equity rather than instead of it, and most venture lenders will not lend without an institutional sponsor on the register, because the credit question they are answering is whether those investors would fund again.

For a company with that syndicate, venture debt is usually the cheapest cash on this list in headline terms, and the cost sits somewhere less visible: the warrant, the draw structure, and the covenants that reference investor support. The warrant is equity handed over in the good outcome, and it is the term most often conceded casually because it costs nothing on the day it is signed. We quote no coverage figure here, because no named source in our set prices warrant coverage as a current market series and an invented range would be worse than silence.

For a company without that syndicate, and that is most of the companies asking the question, venture debt is not expensive. It is unavailable, which is a different problem with a different answer. Bootstrapped and founder-owned businesses with real recurring revenue are looking at the first and third columns of this comparison, and the honest version of the conversation says so early rather than after a quarter of introductions.

When does a recurring-revenue facility replace both, and what changed in 2026?

At scale, and less readily than it did two years ago. A recurring-revenue term facility is sized as a multiple of monthly recurring revenue, usually in the range of 3x to 12x (Ramp, ARR loans explained, Matt Angelosanto, 30 October 2025), runs a three to five year tenor, and then flips from recurring-revenue covenants to conventional earnings-based ones (White & Case Debt Explorer, Ben Ludewig and Evan Rahn, 22 July 2026).

The recalibration matters to anyone choosing today. A sell-off wiped around US$300bn from software equity valuations at the beginning of 2026, private software valuations have contracted between 20% and 35% year on year on one data provider's estimate, and lenders have moved their focus from growth projections to cash runway and profitability, setting a higher bar for what counts as recurring revenue and looking harder at termination risk, churn and customer concentration (White & Case Debt Explorer, 22 July 2026, citing QuantPillar for the valuation contraction).

Pricing moved with it. At the top of the market, borrowers secured recurring-revenue loans at roughly 525 to 550 basis points over the benchmark at loan-to-value ratios of 30% to 35%; margins are now wider and loan-to-value ratios materially tighter (PitchBook data reported by White & Case Debt Explorer, 22 July 2026). The sharpest single reading comes from the placement side: of eight software loans one advisory firm placed in the first four months of 2026, not one was underwritten on annual recurring revenue, and all eight were priced on earnings (Lincoln International data reported by Development Corporate, 19 May 2026).

Read that as a boundary rather than an obituary. The facility still exists and still suits companies with high switching costs, strong retention and revenue a lender can test. But a founder who assumed the recurring-revenue facility was waiting one growth year away should re-test the assumption, because the market that would have written it in 2023 is underwriting on earnings now.

How should the choice actually be made?

In one currency, across three cases. Convert every offer into an annualized cost of the money under the same revenue plan: the revenue-based multiple into an internal rate of return, the venture debt rate plus the modelled value of the warrant, the facility's spread over its benchmark plus fees, and the equity's cost as the value of the stake surrendered at a realistic exit. Then run the same exercise on a downside case and a flat case. The instrument that wins on the base case is frequently not the one that wins on all three.

Then apply two filters the arithmetic will not supply. Availability, which decides whether an option is real: no institutional syndicate means no venture debt, thin recurring revenue means no term facility, and a business that is not yet at the size the facility market serves is choosing among fewer things than it thinks. And reversibility, which decides what a mistake costs: a revenue-based advance ends when the cap is repaid, a facility ends at maturity, and equity does not end.

The last consideration is sequence rather than selection. These instruments form a ladder, and the useful question is often not which one to take but which one to take first. A revenue-based advance that buys the growth which makes a cheaper facility available has done its job, provided it is crossed once and in one direction. Stacking successive advances against the same revenue stream, rather than graduating off them, is the characteristic way this ladder is climbed backwards.