Venture Debt: How It Works, What It Costs, When to Avoid It

The first organic Google result for "venture debt" (US results via Ahrefs, November 10, 2026) is Silicon Valley Bank’s venture debt chapter from the book Venture Deals. It is now published by First Citizens, the bank that took over SVB’s deposits and loans after California regulators closed SVB on March 10, 2023. The chapter warns that when a bank decides it no longer wants to lend venture debt, "it can wreak havoc on your business." It does not mention SVB’s own failure.

Venture debt is a loan, usually a term loan, for startups that have already raised venture capital. Lenders size it against the last equity round, count on the next round to repay it, and usually take warrants on top of interest. Sized small and priced well, it buys months of runway for little dilution. The danger sits in the payment schedule and the default clauses, not the headline rate.

Six of the eight organic results for "venture debt" on that same search are published by banks or financing providers that sell the product or an alternative to it. Startup Yeti sells nothing and has no lender relationship. Below: dated term ranges with named sources, an illustrative all-in cost calculation, four sizing tests, the clauses worth negotiating, and what SVB’s failure means for your loan documents. This is general information, not financial advice. Have your own counsel read any loan agreement before you sign.

How venture debt works

Most US venture lenders are banks or specialty lenders called venture debt funds. A 2020 working paper by Davis, Morse and Wang describes the funds as charging higher rates than banks, using financial covenants less often and demanding more warrants, and found early-stage firms using venture debt in one-third of financing rounds in its pre-2020 sample.

SVB’s chapter states the first rule plainly: venture debt "follows equity; it doesn’t replace it." The lender is underwriting your investors more than your revenue. Venture finance lawyer Andy Hirsch, quoted in the same paper, says lenders to seed and Series A companies study the VCs’ track records and the odds of a Series B, and "that’s really what they are lending against." Law firm Foley Hoag writes that most institutional venture lenders expect at least one preferred equity round, and SVB’s chapter says venture debt "isn’t usually available to seed-stage companies."

Foley Hoag lists three common forms: a single term loan, a delayed-draw term loan released in tranches as you hit milestones, or a borrowing base against receivables, inventory or recurring revenue. This page assumes a term loan, which is paid back in stages:

1. Closing. You pay an upfront fee plus the lender’s expenses, attorneys’ fees included.
2. Interest-only period. Interest but no principal, often for 6 to 12 months per SVB’s chapter (Foley Hoag says anywhere from the first year to 18 months).
3. Amortization. Each monthly payment now includes principal, so the payment jumps. SVB’s chapter says these loans usually have to be repaid within three to four years.
4. Payoff. A final payment (end-of-term or exit fee) is due at the end; paying off early can trigger a prepayment premium.

Some loans skip amortization and repay all the principal at maturity in one bullet payment, which J.P. Morgan says is less common for bank venture lending.

In the lender’s model, the money for all of this comes from your next equity round.

The loan is secured. Foley Hoag describes a first-priority (senior) security interest in all company assets, with intellectual property commonly excluded but kept under a "negative pledge", a promise not to pledge it to anyone else. EisnerAmper instead describes a blanket lien that includes IP, so check which one you are being offered. Existing debt, convertible notes included, typically has to be subordinated and mature after the new loan.

Is venture debt the same as private credit? Not quite. Carta describes private credit as underwritten on EBITDA and collateral, with larger loans and stricter covenants; venture debt is underwritten on the equity behind you. Foley Hoag adds that private credit funds may lend earlier than venture lenders, at a higher interest cost or with bigger equity demands.

The terms, with dated sources

Our research found no independent survey of 2025 to 2026 venture debt pricing, only ranges described by advisors, lawyers, academics and lenders. Read them as descriptions of typical deals, not measured market averages.

A floating rate moves with its benchmark. EisnerAmper quotes a spread over SOFR; J.P. Morgan says some venture loans price off the prime rate instead, which stood at 7.00% on November 6, 2026.

Two ways to count warrants

Warrant coverage is quoted in two different units, and mixing them up can make an offer look several times cheaper or dearer than it is.

• Percent of the loan amount. 5% coverage on a $3 million loan means warrants to buy $150,000 of stock at the strike price. How much of the company that is depends on the valuation.
• Percent of fully diluted equity. The lender can buy a fixed slice of the company, such as EisnerAmper’s 0.5% to 1.5%. One sample term sheet published by Mostly Metrics asked for 0.20% of fully diluted shares, half earned at close and half based on the highest amount drawn.

In the worked example below, 5% loan coverage comes to 0.375% of the company, below EisnerAmper’s whole equity range. Ask which convention a term sheet uses before comparing two of them.

Foley Hoag says warrants are priced either as a penny warrant (a nominal strike, so the lender pays almost nothing to exercise) or at the last preferred round price or latest 409A valuation. It also writes that most venture lenders require an equity kicker, a warrant more often than a right to invest in your next round. Mostly Metrics describes negotiating a fixed success fee instead, such as $200,000 payable on a change in control. So "non-dilutive" oversells it: venture debt is less dilutive than equity, and rarely dilution-free.

The clauses that matter more than the rate

Our overview of startup financing options covers ordinary loans, credit lines and revenue-based financing. A venture loan adds a layer of control those simpler products may not have: consent rights, reporting rights and several ways to call a default.

Negative covenants. Things you cannot do without the lender’s consent. Foley Hoag lists tight restrictions on more debt and liens, investments, acquisitions, distributions and changes to the capital structure, change of control included; J.P. Morgan names dividends, M&A and asset sales. Reporting duties, per Foley Hoag, include monthly or quarterly financials and an annual audit.

Financial covenants. Numbers you must hit. Foley Hoag says they are typically limited to an EBITDA target (often a loss that has to narrow) or a minimum cash balance. Mostly Metrics gives an example of a cumulative revenue covenant tested monthly at about 75% of the board-approved plan. Miss one and, Carta warns, you can face higher rates, less access to credit, or a default.

Material adverse change (MAC). Instead of financial covenants, some loans carry an event of default triggered by a "material adverse change" in your business. Kruze Consulting says a broad MAC lets a lender cite "almost any negative development" as grounds for default, even changes in the wider business environment, and a funding MAC lets it refuse to advance more money on the same grounds. Kruze suggests avoiding the clause, or defining it narrowly with carve-outs for market downturns.

Investor-support default. Foley Hoag describes this as triggered when the lender sees a "clear intention of the borrower’s investors to no longer fund the company" in amounts needed to repay the loan, which leaves "more room for a lender’s discretion" than a financial covenant. Kruze puts it more bluntly: the default can arrive "sometimes regardless of your cash position."

Deposit requirements. Foley Hoag says some institutional venture lenders require you to move most of your deposit accounts to the lender’s bank, though it has become standard to let you keep some funds elsewhere, typically at least two payroll cycles. Accounts outside the lender (and, with a non-bank lender, your accounts generally) need an account control agreement: a three-way agreement among you, your bank and the lender that lets the lender take control of the account after a default.

Cross-default. A default under your loan can trigger defaults under other contracts, and the reverse. SVB’s borrowers found out what that means in March 2023 (more below).

Personal guaranty. Foley Hoag says a lender may ask the founders of an early-stage, pre-revenue company for one: a full backstop, a capped amount, or a narrower "bad acts" guaranty.

What a venture loan really costs: a worked example

This example is illustrative. The assumptions sit inside the sourced ranges above, and the arithmetic is ours.

What it costs in cash:

Run those cash flows as an annual rate and the all-in cost is about 14.5% a year, or 14.0% if you leave out the $25,000 of legal and diligence. A 12% headline rate costs roughly 2 to 2.5 points more once fees count, and that is before warrants.

The warrants: $150,000 of stock at a $40 million post-money price is 0.375% of the company (0.75% at 10% coverage). The equity alternative: raising the same $3 million as extra Series A equity at a $40 million post-money valuation would cost 7.5% of the company, a simplification that ignores the option pool and any change in price.

On ownership math alone, the loan wins easily. Existing shareholders give up 0.375% instead of 7.5%. The 7.125-point difference is worth more than the $924,290 of cash cost once the company’s eventual equity value tops about $13.0 million ($13.7 million at 10% coverage). A company raising $10M at a $40M valuation is betting on being worth far more than that.

So the real cost is somewhere else: the month-13 cash cliff. Say your monthly net burn is an illustrative $400,000. During the interest-only year, the loan costs $30,000 a month, 7.5% of burn. In month 13 the payment becomes $141,220, about 35% of burn, every month for two years. First Citizens, a lender, suggested in September 2023 keeping debt service below 25% of net monthly burn. This loan breaks that guide the day amortization starts, not the day you sign.

The calculation also leaves out two things that cut in opposite directions: the value of the runway the loan buys, and the risk of a default.

When venture debt fits

The sources point the same way: venture debt works as a supplement to a recent priced round, while investors are still committed and the next round has a credible path.

The cleanest case is right after a priced round. The lender follows equity that just arrived, and the interest-only year overlaps a period when burn is fully funded. Real revenue helps too, though not directly: in the lender’s model revenue does not repay the loan, but it makes the next round, which does, more likely.

The backstop case is trickier. A delayed-draw facility you may never use can look like cheap insurance, unless a funding MAC lets the lender refuse the draw. Kruze’s fallback, if you can’t strike the funding MAC, is to draw everything upfront; in our reading that turns insurance into a loan you pay interest on from day one.

One 2020 study of pre-2020 rounds found that startups using venture debt closed less often and exited successfully more often, but its main survival estimate attributes about 60% of the lower closure rate to selection: better companies get the loans.

Four sizing tests worth running against any offer. All are rules of thumb from lenders or advisors, not empirical thresholds.

1. Size against the round. 20% to 40% of the last equity round, or about 6% to 8% of the last valuation (First Citizens, lender, September 2023).
2. Debt service below 25% of net monthly burn (First Citizens, September 2023). In our example, amortization lands at 35%.
3. Organic runway first. First Citizens says the debt should add about six months of runway, and that most lenders expect at least 12 months of organic runway on top of it; Kruze Consulting (updated August 2025) gives the same 12 months.
4. Borrow three to six months of burn. Kruze calls more than three to six months of runway in debt "usually excessive."

When venture debt is dangerous

Pre-revenue, with no institutional backer. Without a VC behind you, the lender’s repayment source is missing. SVB’s chapter says companies without VC investors face significant difficulty getting venture debt at all, and points out that most VCs, unlike most angels, keep reserves for later rounds. Even with an angel-backed profile, it says, significant debt at seed "probably isn’t optimal" when much more equity will be needed. If your backers are mostly angels, our look at the pros and cons of angel investors covers that trade-off. A lender that does say yes may ask for a personal guaranty (Foley Hoag).

Replacing equity to dodge a down round. Wikipedia’s venture debt entry lists "avoid a down round" as a use case. Foley Hoag observes companies lining up facilities "rather than pursue what may be a down-round equity financing," often at their investors’ encouragement. The sources describe how this can go wrong:

• Kruze: if the next round is delayed or comes on bad terms, new investor money may "go straight to paying off lenders."
• SVB’s chapter: VCs "don’t want to see a large portion of their fresh equity used to repay old debt."
• First Citizens, looking back at the 2021 peak: companies that maximized debt now carry debt that "easily tops 20% of their current enterprise value." Its September 2023 guide suggests no more than 10% of durable enterprise value.

None of these sources measures how often the down-round dodge ends badly. Put together, they describe a loan whose repayment depends on the next round, taken because the next round looks shaky.

Burn too high to service. First Citizens gives its own example: a company burning $250,000 a month with a $5 million facility amortizing over 30 months sees its burn rate rise by more than 50% once payments begin. Our check: principal alone is $166,667 a month, two-thirds of that burn, before interest.

How lenders decide

EisnerAmper’s October 2025 guide lists the credit signals lenders weigh: investor quality and syndicate depth; ARR growth, burn multiple and margins; 12 to 18 months of liquidity runway; and team experience and governance. SVB’s chapter says lenders track burn and liquidity to count your months of runway, and ask what your investors ask, including whether more equity will be needed and which metrics will drive the next valuation.

Your burn multiple, net burn divided by net new ARR, is one direct answer to that last question. Lighter Capital, a lender, wrote in 2026 that investors "won’t accept" a multiple "above 2.0x with no story, no trend toward improvement." That matters to the lender too, because its repayment depends on those investors writing the next check.

Venture debt vs revenue-based financing vs equity

Revenue-based financing (RBF) is an alternative for companies with revenue but no VC behind them. Every RBF figure below comes from lenders that sell it: Lighter Capital (RBF page updated February 2026) and Capchase (guide dated 2024 to 2025).

Lighter Capital argues that treating a repayment cap like an APR inflates the cost. It depends on speed. Take Lighter’s own example: $500,000 with a 1.2x cap, so $600,000 repaid. Spread evenly over 36 months, that is about 12.25% a year by our calculation. Repaid in 18 months because revenue grew fast, it is about 23.9%. The fee is fixed; the annual rate is not.

What SVB’s collapse taught borrowers

The facts, from the regulators:

• On March 9, 2023, SVB lost over $40 billion in deposits, and management expected to lose over $100 billion more on March 10. Together, the Federal Reserve’s review says, that was roughly 85% of the bank’s deposit base. At year-end 2022, about 94% of deposits were uninsured, and VC-backed companies accounted for more than half of all its deposits.
• On March 10, the California Department of Financial Protection and Innovation closed the bank and appointed the FDIC as receiver. The FDIC’s release added: "Loan customers should continue to make their payments as usual."
• On March 12, Treasury, the Fed and the FDIC said all depositors would be fully protected, with access to their money from March 13.
• On March 26, the FDIC announced that First-Citizens Bank & Trust would assume all deposits and loans of the bridge bank; branches opened as First-Citizens on March 27.

The Fed’s review spells out how SVB’s lending and banking were tied: "SVB’s loan agreements with certain borrowers required them to use other services of SVB or an SVB affiliate, including maintaining their primary operating deposit accounts with SVB." The agreements did not bar borrowers from opening accounts elsewhere.

Law firm Wilson Sonsini explained the bind in a March 14, 2023 client alert (reprinted by EACCNY). As the firm described them, SVB loan agreements typically required borrowers to keep all or part of their funds at SVB, with account control agreements on most outside accounts, and some liquidity covenants counted deposits held at SVB. Moving money out could breach those covenants, let the lender demand the whole loan back, and set off cross-defaults under other contracts. Replacement financing on similar terms, the firm warned, could be hard to find.

Depositors, borrowers among them, were made whole by a decision announced two days after the closure. In our reading, nobody could have counted on that in advance, and the lesson is concentration: when your lender and your bank are the same institution, a problem at the bank becomes a problem in your loan agreement.

Since then, Mostly Metrics called the deposit requirement "a huge sticking point post-SVB" in February 2024 and suggested negotiating for 50% or less held with the lender. EisnerAmper, in October 2025, described "non-bank funds stepping in, underwriting tightening, and pricing moving upward."

The venture debt market in 2026

Through September 30, 2026, VC-backed US companies took on $81.7 billion of debt in 499 deals, according to the PitchBook-NVCA Venture Monitor (Q3 2026 edition), already more than the $69.7 billion of all of 2025.

That is not the market a Series A founder walks into. It works out to about $164 million a deal, and it is mostly tech: $76.4 billion across 446 deals, in a PitchBook chart titled "AI leading to high loan value for tech." A second chart, of deal value by stage, is titled "Frontier model loans leading to high venture-growth loan value." The dataset counts all debt products raised by VC-backed companies at any stage, so very large AI facilities move the total. PitchBook’s median early-stage loan (Series A and B) was $3.0 million in 2025 and $11.0 million in 2026 so far, against averages of $62.3 million and $106.3 million.

Questions to ask a lender before you sign

Each comes from a risk above. Whatever the answers, get them written into the loan documents.

1. How have you treated borrowers in a down cycle, and have you been tested by one? (First Citizens suggests asking this about any lender.)
2. Is there a MAC event of default or a funding MAC, how is it defined, and will you carve out market-wide downturns?
3. Is there an investor-support default, and what would count as evidence our investors have stopped backing us?
4. What conditions apply to each draw of undrawn money?
5. What share of our deposits must sit with you?
6. Which of our other agreements would a default here cross-default into, and the reverse?
7. Which financial covenants apply, how often are they tested, and do any start only after we draw a certain amount?
8. Are warrants a percentage of the loan or of fully diluted equity, what is the strike, and are they earned at close or on draw?
9. What are the upfront, legal, monitoring and final fees, and what is the prepayment premium schedule?
10. Will you ask for a personal guaranty, and is it full, capped or bad-acts only?

Before signing, put the offer through the month-13 test: work out the amortizing payment, divide it by your net burn, and see what is left of the runway you were buying.