Article · Quantum metrics
The five numbers a quantum founder must defend in two minutes.
A quantum partner asks for five numbers in every first call. They are not the five on your homepage. They are the five on the napkin. If you cannot defend each in 30 seconds, you do not defend them at all.
A quantum VC sits down, opens your model, and asks five questions. Each question is a probe at one specific number. The founder who has prepared answers each in 30 seconds, with the methodology behind it. The founder who has not, talks for ten minutes and ends up explaining why the gross margin "looks weird".
The problem is not the numbers themselves. Most quantum founders have the five somewhere in the model. The problem is they confuse bookings with pipeline, present a gross margin that quietly includes grant offset, or show runway with 100 percent of expected grants landing on time. When the VC asks the second-order question ("strip out the grants, what does it look like?"), the founder freezes.
Five numbers, three mis-statements per number on average, the same patterns every time. Here are the five and how to defend them.
01Bookings
What did you actually sell?
How a VC asks
What is your bookings number, year-to-date?
Show me the signed contract value.
Distinguish bookings from pipeline.
Which contracts are PO-backed?
The number, defined
Bookings equals the total dollar value of signed customer contracts in the period. Single year, not multi-year aggregate. Each booking has a PO number or signed master agreement with a milestone schedule. Anything not signed is pipeline and goes in a different column. For a quantum company at Series A, bookings is small and lumpy; that is fine, as long as the few that exist are real.
How founders mis-state it
Three classic mis-statements. One: counting unsigned MOUs as bookings. MOUs are pipeline at best, marketing at worst. Two: counting LOIs (letters of intent) without commitment language as bookings. Three: counting a multi-year contract as a single-year booking, which artificially inflates the year. The VC has read 200 quantum decks; they sort it out faster than you would like.
How to defend it on a napkin
A one-page bookings register listing each customer, contract value, signed date, PO number, and milestone schedule. Sum at the bottom equals the bookings number you defend. Anything not in this register is not a booking. The VC will accept "we have $480k in pipeline" if it is honest; they will not accept inflated bookings.
02Backlog
What is left to deliver, and when?
How a VC asks
What is your backlog?
How much of bookings is recognized as revenue already?
When do the deliverables hit?
What is the cash conversion cycle on the backlog?
The number, defined
Backlog equals signed bookings minus revenue already recognized. It is the revenue you have a right to but have not yet earned. For milestone-based quantum POCs, backlog is often large relative to recognized revenue, and the timing depends on the technical milestone calendar (when chip v4 ships, when the benchmark is met, when the customer signs acceptance).
How founders mis-state it
Three classic mis-statements. One: confusing booking value with backlog. Backlog is what is unfulfilled, not the contract total. Two: omitting the milestone calendar so the investor cannot see WHEN backlog converts. A $300k backlog that lands in Q3 reads differently from one that lands in Q1 next year. Three: double-counting backlog in pipeline (the same dollar shows up twice).
How to defend it on a napkin
A backlog table, contract by contract: milestone schedule, expected recognition month, cash collection month. The VC wants to see the next 12 to 18 months of conversion to revenue. If chip v4 ships in Q1 2027 and that triggers a $180k milestone payment, that line is on the table with the date. The day chip v4 slips, the date moves; that is what the milestone calendar is for.
03Gross margin (with and without grant offset)
How much of the unit economics is real?
How a VC asks
What is your gross margin?
Strip out the grant offset, what does it look like?
What is the cost of revenue on a paid POC?
Gross margin at scale, how does it look?
The number, defined
Gross margin equals revenue minus direct cost of revenue, divided by revenue. For quantum POCs, direct cost of revenue includes scientist time on the project, fab time, control electronics depreciation allocated to the contract, foundry-slot amortization, and customer-specific cloud compute. Two versions matter: with grant offset (cash reality today) and without (unit economics at scale). Investors want both.
How founders mis-state it
Three classic mis-statements. One: subtracting non-dilutive grants from cost of revenue, which artificially boosts gross margin. This is the most common error and FDD will catch it. Two: attributing too much fab capex to general operations instead of cost of revenue, which under-states the unit cost. Three: showing only "best case" gross margin from one POC and projecting it to all customers. Quantum is bespoke at Series A; one good POC is not the steady state.
How to defend it on a napkin
Two gross margin numbers, side by side: with grant offset and without. Founders who only show the "with" number lose credibility; founders who only show the "without" leave money on the table because they hide the cash reality. Show both, label them clearly, and explain the trajectory: today the "with" is X percent and the "without" is Y; at scale (5 paying systems, no grant subsidy on the unit), both converge near Z. Calibrated honesty closes more rounds than optimism.
04Capex per quantum system
How much hardware does it take to make a unit of revenue?
How a VC asks
What is the capex per system?
What does it cost to build the next chip?
How does capex scale with revenue?
What is the depreciation profile?
The number, defined
Capex per quantum system equals the total dollar cost to deploy one new system. Cryostat or dilution refrigerator, control electronics, lasers (for atom systems) or photonic packaging, lab build-out per system, foundry campaign cost. Plus the depreciation schedule by component: typically 7 to 10 years for cryogenic systems, 3 to 5 years for control electronics, 5 to 7 years for lasers and optics. Investors want the unit number and the schedule, separately.
How founders mis-state it
Three classic mis-statements. One: combining R&D capex (the chip-design hardware that does not ship to customers) with production capex (the systems that do) into one number, which inflates the unit cost. Two: showing only "incremental" capex without the upfront fab investment, which under-states the cost of the first system. Three: expensing capex at purchase instead of capitalizing it, which we covered in the FDD prep checklist; it breaks both gross margin and EBITDA at the same time.
How to defend it on a napkin
Capex per system at current state and at scale, with the components broken out: cryostat (about $400k), control electronics (about $150k), fab campaign per chip-batch (about $300k amortized over the units yielded), packaging (about $50k). The depreciation schedule by component on a separate line. Show how it scales: at 5 systems, fab amortization per system drops by 60 percent; at 10 systems, the cryostat backup unit becomes a shared resource. The ratio capex-to-revenue is what the investor underwrites.
05Runway sensitivity
How much runway do you actually have, under different scenarios?
How a VC asks
How much runway do you have?
What is the runway if the next grant does not come through?
What is the runway if you sign two more POCs?
When do you raise next?
The number, defined
Runway equals cash on hand divided by monthly burn. For quantum, both numbers are unstable. Cash fluctuates with grant draws and POC milestone payments. Burn varies with hire timing and capex spending. A single runway number is meaningless without sensitivity around the inputs. Quantum-fluent investors expect three scenarios on the same line: base, downside, upside.
How founders mis-state it
Three classic mis-statements. One: showing runway with 100 percent of expected grants landing on time, which rarely happens. Two: showing runway at current burn without the planned hires that the round funds (which under-states future burn). Three: skipping the capex spend line entirely, which moves the cliff. Combined, an optimistic runway can over-state the actual months by 4 to 6.
How to defend it on a napkin
Three scenarios at the bottom of the model. Base: 50 percent probability that expected grants land on schedule, current hire plan, current capex commitments. Downside: no grants land in the period, hires delayed by one quarter, capex on track. Upside: grants land plus 2 POCs sign plus hire on time. Show all three with the implied raise date for each. The investor will run their own sensitivity; you might as well show yours and own the calibration. The founder who pre-empts the question wins the conversation.
Five numbers. Defend each in 30 seconds. Total: two and a half minutes, a margin under the two-minute standard the investor sets. The pattern repeats whether the partner meeting is in Boston, Paris, Munich or Singapore.
The compounding test: if the founder gets bookings, backlog, gross margin, capex per system and runway sensitivity right, the rest of the call is about the technology. If the founder gets one wrong, the rest of the call is about the model, and the technology never gets the airtime it deserves. The cost of preparation is one afternoon with the model; the cost of skipping it is the conversation never reaching the science.
Build the napkin. The napkin is the model. The model is the deck. The deck is the round.
Sanity check
Want a 2-minute audit of how you defend these five numbers?
The Fundraise Readiness audit grades your bookings, backlog, gross margin, capex per system and runway sensitivity the way a quantum VC reads them. Score, three specific gaps, plan in your inbox.