Reading a founder's burn rate.

Burn isn't one number — it's three. The one in the deck is usually the prettiest. Here's how to triangulate the real one — and what it tells you about runway.

01The three burns

Founders quote burn the way that flatters them. There are three different numbers in play in any growth-stage company, and the deck usually shows the smallest. Each is defensible in isolation. The mistake — the one that makes IC conversations talk past each other — is comparing one company's adjusted burn to another's net burn and concluding the businesses have similar economics. They don't.

The three numbers, in order from largest to smallest:

• Gross burn. Total cash out, before any cash in. The headline operating cost. This is what payroll, rent, infrastructure, and contractor invoices add up to in a quarter. It does not net out revenue, financing inflows, or one-time payments. It is the cleanest measure of how expensive the business is to run at its current size.
• Net burn. Cash out minus cash in from operations. The actual rate at which the bank balance is declining. This is the number that determines runway. It nets revenue against operating cost. It does not net out financing.
• Adjusted burn. Net burn minus “one-time” or “non-recurring” items the founder prefers to exclude. A model fine-tune that cost $200K. A trade-show booth that ran $80K. A two-month consultant on an org redesign. Each is, by itself, plausibly excludable. Together, they are how a $640K/month net burn becomes a “$420K/month adjusted run rate” on slide fourteen.

The deck almost always quotes the third. The IC almost always treats it as the second. The actual cash position behaves like the first. This is how companies look fine in the partner meeting and run out of cash three months later.

02How to triangulate the real number

You can reconstruct net burn from three data points the founder will share without protest:

1. Cash position at the start of the quarter
2. Cash position at the end of the quarter
3. New financing or extraordinary inflow during the quarter

Net burn = (start − end + financing inflows) ÷ 3. The formula is deliberately rough. It catches all the “adjustments” the deck has already made because it does not depend on what was or wasn't counted as recurring. It is the cash that actually left the building.

Ask the founder to confirm the three numbers against their bank statement, not their accounting export. The bank statement is the ground truth. The accounting export reflects whatever conventions the founder's controller has applied — which can be reasonable, but they introduce the kind of definitional drift that exactly this exercise is designed to surface.

03The runway math

A company with $4M in the bank and a $400K net burn has ten months of runway. The deck will probably claim fourteen, because the deck is using adjusted burn and assuming a revenue ramp that hasn't happened yet. The difference between ten months and fourteen months is the difference between a healthy round and a fire sale.

Three months matters because financing rounds take three months to close from term sheet, sometimes longer. Founders raising at month ten of runway are negotiating from strength — they have six months of cushion past the close, time to walk from a bad term sheet, and the optionality to pick the right investor over the fastest one. Founders raising at month five are taking the first term sheet that arrives because they cannot afford to wait. The terms reflect this. So does the cap table.

This is also the reason “we have eighteen months of runway” is the safest answer a founder can give in a partner meeting. It implies they are raising on plan, not on emergency. The unsafe answer is more interesting: “Twelve, but we're planning to extend it through Q3.” That implies the founder is operating with a margin of error, knows it, and is making active choices. The follow-up is: which choices? Headcount freeze? Cost-cutting? Revenue acceleration?

04Burn multiple — the cohort comparison

The most useful single number for comparing burn across companies is burn multiple — net burn divided by net new ARR for the same period. Coined by David Sacks at Craft Ventures, it has become the institutional shorthand for capital efficiency at growth-stage. Here is how to read it:

• Burn multiple < 1.0. Best in class. The company is acquiring revenue faster than it's burning cash. This is rare and is usually associated with product-led growth businesses or very high-margin SaaS with a low CAC.
• 1.0 – 2.0. Healthy growth-stage company. Most institutional-quality businesses sit here. The company is investing aggressively but with measurable return.
• 2.0 – 3.0. Typical for early-stage with heavy R&D investment. Acceptable if the founder can articulate why the multiple will improve as the company scales — usually a function of fixed costs being amortised across more revenue.
• 3.0+. Ask why. Sometimes structural — long sales cycle, deep-tech, regulated category. Sometimes a signal that the company is buying revenue at unsustainable rates and the burn multiple will not improve without a material change in the unit economics.

The cohort median for SaaS Series B is currently around 1.4. Vertical SaaS skews slightly higher because of higher CAC. Applied-AI startups skew slightly lower because of higher gross margin on inference-light workloads. We benchmark every dossier against the matched cohort, not the universal median — comparing a vertical-SaaS company at 2.1 against the universal SaaS median is the kind of category error that gets the IC to the wrong answer.

05What the founder won't volunteer

Three structural issues hide inside any burn number, and the founder will almost never volunteer them. Each is worth asking explicitly.

• Founder compensation. Does the burn include founder salaries at market rate, or compressed founder salaries that will need to be raised post-round? A company quoting $400K/month burn while paying its founders $80K/year is implicitly understating burn by $30K – $40K/month per founder once the round closes. The founders deserve market comp; the IC deserves to know that's in the model.
• Amortised one-time costs. Was the model fine-tune a one-time cost that's done, or does it need to be redone every six months as the data drifts? Is the trade-show budget actually one trade show, or four trade shows the founder is calling one-time because none of them are recurring in the same quarter? The answer reveals whether the “adjusted” framing is honest accounting or a quiet pattern of moving the goalposts.
• Baked-in headcount expansion. Does the current burn assume the current team stays, or are new hires already in the model for the next two quarters? A burn rate that assumes the team grows from twelve to twenty looks identical to a burn rate that assumes twelve people stay. The two companies have very different cash trajectories.

06How we surface this in the dossier

Every Investment Committee dossier we generate reconstructs burn from the deck's claimed numbers and the inferable balance-sheet movements, then compares the reconstructed figure against the founder's claim. The variance is quoted explicitly in section four, Financial Forensics. Where the variance is material — more than 15% — it surfaces as a kill flag if the founder has not already disclosed the framing. We don't treat every variance as bad faith. Most founders genuinely do not realise their accounting convention differs from the institutional one. The point is to surface the gap before the term sheet, not after.

The single discipline that catches the most issues, by a wide margin, is the three-data-point reconstruction above. It takes five minutes per company. It catches roughly half of all founder-vs-reconstructed variances we eventually flag. The remaining half require comparing against cohort benchmarks or against the founder's own historical numbers. But the first half is portable and free, and any IC partner can run it without specialist tools. We recommend it as the floor of burn diligence — the minimum that should happen on every deal.