Flow Capital’s Managing Director, Josh Axler, joined Ben Murray on The SaaS CFO Podcast to talk about how lenders underwrite SaaS companies for venture debt. This post pulls together the key points from that conversation.
Hitting $2.5 million in ARR gets you a meeting. What gets you a term sheet is everything underneath that number: whether the revenue is live or just signed, whether customers renew, and whether your plan holds up if things go slower than expected.
There's a practical upside to knowing all this. Founders who understand what lenders will ask can have the answers ready before the process starts, and preparation, more than anything else, is what decides how fast a deal closes.
Here are the key takeaways from the conversation. You can listen to the full episode here.
Almost every conversation with a lender starts the same way: what's your ARR?
At Flow, the entry point is around $2.5 million in ARR, usually where product-market fit starts to show. From there, loan sizes of $1 million to $15 million map to companies between roughly $2 and $20 million in ARR, so that's the range where venture debt makes the most sense.
The headline number is just the starting point, though. The first thing a lender wants to know is what's in it. A company might report $3.5 million of ARR when part of it is signed but not yet live: customers are still being onboarded and haven't started paying. That's not a problem in itself, but the difference matters, and it should be clear from the first conversation.
From there, the questions get more specific. How spread out is the customer base? How long have customers been around, and have they renewed at least once? Are they buying more over time, whether that's seats, usage, or features? Size is easy to read off the headline number. Judging how solid the revenue is takes all of the above, and that's where lenders spend most of their time.
Founders can tell a lender that customers love the product, but retention is the proof. If customers keep paying, and ideally pay more over time, they're voting with their wallets. That's why gross and net revenue retention carry so much weight in SaaS lending.
Right now, below roughly 80% GRR or 100% NRR, deals get much harder to do. Part of the reason is AI. With so much uncertainty about how durable software revenue will be, a company already losing customers is more likely to see that get worse, not better.
The aggregate number isn't always the full story, though. Take a company with GRR in the low 80s. Weak on the surface. Split the customers into segments, and one group retains at nearly 100% while the other sits at 70%. If management has spotted the difference, is walking away from the weak segment, and is putting its growth dollars behind the one that works, the story changes. Some lenders will run the aggregate number through a screener and stop there. Better ones will look a level deeper, but only if the company has done the work and can explain it clearly.
The practical move: if you have a decent-sized customer base, segment your retention data before a lender asks you to.
Retention shows the revenue is solid. The next question is what it costs to grow it.
Growth matters, but context matters more. Ten percent growth on $2 million of ARR is a different business than 10% on $20 million; the second one is adding $2 million a year. A useful rule of thumb: if you're considering venture debt, your revenue should be growing faster than the interest rate on the debt.
From there, the analysis works down the P&L:
The biggest misalignment between founders and lenders is the forecast. Founders are used to pitching the growth case: the model built for the board or for equity investors, hockey stick up and to the right, everything going perfectly the moment the money lands.
Lenders need to see the range: a growth case, a base case, and a downside case, plus a clear view of what management can do if the plan slips. Can hiring slow down? Can spending come down? How much runway is left if sales land later than expected? Debt has to be repaid whether or not you hit the board plan, so the downside case isn't pessimism. It's the whole point.
The assumptions matter as much as the model. If the money is funding four new salespeople, expect questions like: How many reps do you have today? How many hit quota last year? What does your pipeline look like? Are the new hires being interviewed now, and do their start dates in the model match reality? The closer the assumptions are to what's happening in the business, the more believable the forecast. And believability is what gets funded.
When a financing process goes sideways, it's usually not the business. It's that the company can't find the data a lender needs, or can't present it in a way that shows they're on top of their numbers. Fixing that in advance is the single biggest thing a founder can do to speed up a raise.
It starts on the first call. Whoever's on the phone should be able to answer the basics without going away to check: What's your current ARR? What are you burning, roughly? How much capital do you need? How much equity have you raised, and at what valuation? These should roll off the tongue.
Then the data room has to back it up. When what management says lines up with what's in the documents, trust builds. When it doesn't, trust breaks, and it's hard to win back mid-process.
A complete diligence package usually includes:
Not every early-stage company has all of this ready, and that's fine, as long as the underlying pieces exist to build it, whether in-house or with a fractional CFO.
Either way, the timing advice is the same: start at least six months before you need the money. Time kills deals. But when you're preparing your business, time is your friend.
AI has added a layer to lending that didn't exist a few years ago. The core question is blunt: is this business a product or a feature? And what does that mean when competing products can be built fast, not just by new entrants but by your own customers?
Lenders are also watching how companies use AI internally, and what AI costs will do to margins over time. Those costs are cheap today, but companies blowing through their annual AI budgets by mid-year is likely to become the norm.
The bigger shift is on the revenue side. AI-native companies are hitting $2.5 million in ARR within months of founding. The growth can be incredible and the lending decision nearly impossible: a business growing 100% month over month while losing 40% of customers month over month might turn into a great company, but lending to it at that stage is very hard, because there's no proof yet that the revenue survives a renewal.
In other words, the age of your ARR now matters as much as the amount.
Usage-based revenue follows the same logic. With enough history (years of usage data, repeatable patterns, strong retention), a lender can treat it much like subscription revenue. The other test is value: if customers clearly get many times what they pay in savings or new revenue, that revenue is much easier to trust.
Being ready covers how you present the business. The other half of the conversation is what the money is for.
Founders turn to venture debt for a simple reason: they need growth capital, they believe the value of the business is going up, and they'd rather not sell equity cheap today. Three uses come up most often.
What doesn't work: using debt to plug a hole in a business with no path to profitability or a future round. That's equity risk, not debt risk. Most venture debt borrowers are still burning cash; that's normal. But there has to be a realistic path to profitability, or a credible plan to raise equity, within the term of the loan.
There's no formula, but there are reference points. Flow can go up to roughly 1x ARR, checked against loan-to-value in the 10 to 20% range. Think a $2 million loan to a business worth $20 million on its last raise.
Growth changes the math. A company growing 100% that borrows 1x ARR today is only at 0.5x a year later; fast growers shrink their own leverage. Flexibility helps too: a company spending heavily on growth that it could pull back in a rough patch is a safer bet than one with fixed costs it can't touch. And the end goal matters: is the company heading toward cash-flow positive, a next round, or a sale? How much you can borrow depends on how you'll eventually repay it.
Ben Murray closed the podcast with a lightning round. Here are Josh's answers:
The best time to become diligence-ready is before you need the financing. If you're a founder or CFO thinking about growth capital, listen to the full episode, or reach out to Josh on LinkedIn. He answers.
