Article · Quantum diligence
Why quantum rounds die in diligence, not in the pitch.
Quantum founders nail the partner meeting, then spend six weeks answering the same three questions in the data room. Not because the deck was wrong. Because the books are.
Quantum diligence
7 min read
Published Q2 2026
Gaetan Brillaud · CFO Fractal
A quantum partner meeting goes well. The investor likes the team, the science, the wedge. The week after, you get a term sheet, and the cap table starts to look real. Then financial diligence opens, and the round dies in the data room.
The partner meeting is the easy part. The investor wanted to like you and you helped them. Then the financial diligence team gets access to the books, and the angle of the conversation flips. They are paid to find things to ask about, not to like you. The clock starts on a 30-day or 45-day exclusivity, and every gap they raise is a new ticket on the closing path.
After hundreds of quantum data rooms, three gaps show up in eight out of ten cases. Three gaps where SaaS playbooks are wrong and quantum-fluent investors know it. Fix them and diligence shortens by three weeks; ignore them and the round re-prices or walks. Here are the three.
01Freedom-to-operate
Who actually owns what makes you valuable?
What surfaces in diligence
Walk us through the patent estate.
Who owns the IP from the founders' time at the lab or university?
Have you done a freedom-to-operate analysis?
Is anyone licensed under your patents already, formally or informally?
The underlying gap
Quantum hardware sits on top of dense academic IP, often co-owned with universities, national labs, or thesis advisors under terms founders read once at incorporation and filed away. The diligence team's nightmare is funding you to a $200M valuation and discovering the spin-out license is non-exclusive, capped, or revocable on a six-month notice. They will read every page of every license. They are looking for the one clause that lets the lab pull the rug.
Where founders trip
"We have twelve patents pending" is the wrong answer. The right answer is the FTO landscape: of the patents that matter for our architecture, three are ours, four are public, two are spun out under an exclusive license we have reviewed (here are the terms), and three are held by [larger company] but the claims do not read on our implementation because [specific reason]. If you cannot speak to the landscape that way, the diligence reviewer concludes you have not done the work, and the legal team will ask for a $40k FTO opinion before the round can close. That is three weeks added to the timeline, often past the term-sheet expiration.
The fix
Get the FTO summary done six weeks before the term sheet, not after. License agreements with any incubator university or national lab live in /05_IP and /01_Corporate, both. Tech-transfer royalties owed to the lab are accrued in the period earned, even if not yet paid. Patent counts are vanity. Patent landscape is the moat. Know which one you have, and never confuse them.
02Milestone revenue recognition
When did the money actually count?
What surfaces in diligence
Walk us through the bookings-to-revenue waterfall.
What's your revenue recognition policy on the bank POC?
Is the recognition cash-basis or accrual?
Show us the customer acceptance for milestone three of the EUR 180k contract.
The underlying gap
Quantum POCs are typically milestone-based, not subscription, not point-in-time at delivery. ASC 606 and IFRS 15 have explicit guidance for either model, but the policy has to match the contract. Most quantum founders default to cash basis (recognize when paid) because it is what the bank account shows. Cash basis on a milestone POC is a guaranteed FDD finding and a quality-of-earnings adjustment that lowers the recognized revenue used to underwrite the round. The valuation moves accordingly.
Where founders trip
"We recognize when the customer pays" is not a policy. Neither is recognizing the entire POC at signing. Both are common, both fail FDD. The two acceptable models are over-time (proportional to work performed) or point-in-time at acceptance of each technical milestone. Pick one, document it on a one-page memo, apply it consistently to every contract. The day you let the bookkeeper improvise is the day the FDD reviewer flags it.
The fix
A revenue recognition memo, one page, on file in /03_Financials before the term sheet, covering each revenue stream: paid POC, hardware sale, cloud subscription, co-developed solutions, IP licensing. A monthly bookings-to-revenue waterfall that ties to the bank statement and the customer invoice register. Refundable or contingent revenue (POC failure clauses, performance penalties tied to fidelity targets) held in deferred revenue until the contingency expires. Get this right at incorporation, not at term sheet.
03Non-dilutive grants flow
Half your runway is non-dilutive. Where does it sit on the books?
What surfaces in diligence
Show us the grant ledger, by program.
How are SR&ED, IRAP, and CIR claims treated, refundable tax credit or grant income?
Is the EIC blended (grant plus equity) split correctly between the two components?
What's the FX policy on the DARPA receipts in EUR books?
The underlying gap
A quantum company at Series A often runs four to six active programs across three or more jurisdictions: SR&ED Canada, IRAP Canada, CIR France, EIC Accelerator, France 2030 / PEPR Quantique, DARPA US2QC / QBI / ONISQ, ARIA, DOE, NSF, EU Quantum Flagship, EuroHPC. Each treats the books differently. Each requires reporting. Lumping them into a single "Other income" line on the P&L is the easiest finding the diligence team will make, and the one that costs the most.
Where founders trip
Three classic mis-classifications, in order of frequency. One: Bpifrance "avance remboursable" booked as grant income instead of debt. Two: EIC blended (the equity part of EIC Accelerator) booked as grant when it converts to equity. Three: SR&ED refundable tax credit booked as "Other income" when it should reduce R&D expense. Each shifts EBITDA in the wrong direction; combined, they are why the diligence team's normalized burn calculation differs from yours by 30%, every time.
The fix
A separate sub-ledger per program with the original agreement, milestones, draws, and reporting obligations. Each grant classified correctly: forgivable loan vs grant income vs refundable tax credit vs investment grant vs repayable advance. FX policy documented for cross-border grants (USD DARPA in EUR books, etc.). And the kicker: grant-tied milestones tagged in the financial model and the burn forecast, not just in the legal folder. If a grant draw is the difference between 12 and 18 months of runway, the diligence reviewer needs to see the cash-timing assumption to underwrite the round.
Three gaps. The same three, every quantum data room, eight times out of ten. The pattern repeats whether the round is in Boston, Paris, Munich or Singapore.
Founders read this and conclude the answer is to hire a CFO before raising. The actual answer is smaller and less expensive: spend $5k to $15k on an external accountant or fractional CFO who reviews the books for these three things, six weeks before the term sheet. The cleanup work is bounded; the cost of skipping it is unbounded. A failed $200k QoE adjustment shifts your valuation by $4M at 20x forward earnings, every time.
Fix the books before you fix the deck. The deck got you the meeting. The data room gets you the round.
Sanity check
Want a 2-minute audit of the three gaps before your data room opens?
The Fundraise Readiness audit grades the three gaps that kill quantum rounds in diligence: freedom-to-operate, milestone revenue recognition, and grant accounting. Score, three specific gaps, plan in your inbox.