How the Ask Falls Out of the Financial Model: Raise Size, Use of Funds, and Runway

Most founders build their fundraising ask backwards. They decide how much to raise, usually by glancing at what similar startups raised, and then reverse-engineer a spreadsheet that justifies the number. Investors have seen that move a thousand times, and it shows.

A good model works the other way round. The raise size, the use of funds, and the runway aren't inputs you feed the model; they're outputs the model produces once you fix one thing: the milestone you need to hit before the next round. Get that inversion right and the number stops being something you defend after the fact. It was never arbitrary to begin with.

This is where a financial model earns its keep as a communication tool rather than a forecast. An investor's first two questions are almost always "how much are you raising?" and "what does it get you?" When both answers fall directly out of the same model that produced your revenue build and your burn, same assumptions, same drivers, no separate story, you've demonstrated the thing diligence is really testing for: that you understand your own business well enough to have earned the ask.

A note on the numbers throughout: these are 2025-2026 benchmarks drawn from primary investor and market sources (Y Combinator, a16z, Carta, and European VC data) rather than the advisory-blog consensus that tends to repeat itself. Where those sources disagree, and they do, I've shown the range rather than pretending there's a single right figure. Everything is euro-denominated and written with a European founder in mind, where valuations and round sizes run meaningfully below the US headlines.

Start from the milestone, not the dollar amount

The most common mistake is raising for a specific dollar amount rather than a specific timeframe. "€2M because that's the going rate" and "€750K because we want low dilution" are both guesses, and investors can hear the difference. The disciplined version runs the other way: decide what you need to prove before the next round, then let the model tell you what proving it costs.

The canonical version of this logic comes from Paul Graham's "default alive or default dead" framing. The essential question is whether your startup, on its current trajectory and spend, reaches profitability before the money runs out. Graham's warning is the fatal pinch: default dead, slow growth, and not enough time left to fix it - the trap founders fall into when they raise for calendar time and hire ahead of a growth that never arrives. Capital, in this frame, is fuel for something specific. Runway is the consequence of the milestone you're funding, not the goal itself.

This also resolves a tension you'll see across fundraising advice, where "raise for eighteen months" and "don't raise for runway, raise for milestones" sit side by side. They're the same rule stated from two ends: the milestone sets what you must achieve, the burn required to achieve it sets the duration, and the runway is simply that duration made visible.

So the sequence is:

1. Name the milestone that makes you fundable at the next stage. At seed this is increasingly revenue-defined, roughly €1M+ ARR with strong retention, and year-on-year growth in the 2-3x range, has become a common working baseline for Series A readiness (though the exact bar varies widely by sector, and AI companies are held to a different standard entirely).

2. Read the monthly burn required to reach it, off your hiring plan and spend assumptions, the model already contains this.

3. Add a fundraising buffer so you aren't raising the next round on fumes.

4. The product is your raise size.

The formula, and the buffer that has grown

The formula is refreshingly simple:

(months to milestone + fundraising buffer) × monthly burn = raise size

A worked example on a European cost base: a four-person Berlin team burning roughly €55-65K per month, targeting a milestone twelve months out, with a six-month buffer, needs about (12 + 6) × €60K ≈ €1.1M, so you'd raise around €1.2M. Build this off your hiring plan, not any published example; European early-stage burn for a comparable team typically runs below the US figures you'll see quoted.

The buffer is the part founders shortchange, and it's the part that has moved most - which is where the primary data matters, and where you should resist quoting a single tidy number. The gap between rounds has stretched substantially, but by how much depends on which cut you read:

• Carta's Q2 2025 data put the median seed-to-Series-A interval at about 616 days, a little over twenty months, and more than two months longer than two years earlier.

• Other Carta cuts are harsher: the Q4 2024 median was reported at around 774 days (~26 months), up 84% since 2021.

• And in a Carta analysis of thousands of startups, 39% of companies that raised a Series A in Q3 2025 took three or more years to get there, roughly double the rate of 2019.

The exact median depends on the quarter and the cut, so the honest statement is directional: the typical gap has pushed past two years, and a large and growing minority now take three-plus. Quoting "616 days" alone is the rosiest reading of a harsher picture.

Against that backdrop, how much runway should you raise for? The primary sources genuinely disagree, and the spread is the answer:

• Y Combinator advises raising for 12-18 months of runway, with typical seed rounds of roughly $750K-$2M at 20-25% dilution. That's the lean end - enough to build, test, sell, and show traction without living in fear of zero cash.

• Market conditions - the stretched gap between rounds - push some advisors toward 18-24 months or, at the cautious extreme, planning for a full 24-plus.

Neither is "correct." YC's number reflects a bias toward staying lean and default-alive; the longer-runway camp reflects the reality that the next round now takes longer to reach. Pick deliberately, and know which trade-off you're making: more runway buys safety at the cost of more dilution now.

Whatever number you choose, the buffer exists so you begin the next raise from leverage rather than desperation. The working rule: start fundraising with roughly nine or more months of runway left, so a raise that drags doesn't force bad terms.

The dilution constraint that can override the whole calculation

Here's where the model meets a wall it can't compute its way past. The milestone maths produces a raise size, but that raise implies a dilution, and dilution has a ceiling the market sets, not you.

On a post-money SAFE the maths is direct: dilution equals amount raised ÷ post-money cap. Raise €1M on a €10M cap and you sell 10%. The reason so many founders cluster near 20% at seed is that round sizes and caps have risen roughly in proportion, so the ratio holds even as the euro figures climb. Globally, Carta's seed and Series A dilution medians both sit around 19-20%, near long-run norms.

For a European raise specifically, the valuation base is lower than the US headlines suggest, which tightens the ceiling. The sources don't perfectly agree, which is itself the point:

• European pre-seed/seed median pre-money valuations were roughly €5M in Q3 2025 (France/Benelux around €5.0M, the UK around €4.7M).

• European seed rounds typically run €1-2.5M, at valuations roughly 20-30% below comparable US companies, with dilution averaging around 21%.

• A broader European guide cites pre-seed around €1-3M pre-money and seed around €5-12M, with valuations 30-40% below 2021 highs.

Treat these as overlapping bands, not fixed rates. They're shaped by sector, country, and cycle. AI is the largest distortion: AI startups command materially larger rounds and higher valuations than the broader market, so any headline median blends two very different populations.

The practical move when your milestone-driven raise collides with the dilution ceiling is to flip the calculation: pick the dilution you can accept, then back into the maximum raise that valuation supports. And when the two numbers don't meet, that gap is information, not an annoyance to model around. It usually means one of three things - the milestone is too expensive for this stage, the valuation expectation is off, or you should be raising a smaller bridge to a nearer milestone instead of a full round. The model surfaces the tension; strategy resolves it.

Use of funds is just the model, disaggregated

"Use of funds" sounds like a separate slide, but it isn't a separate exercise - it's your burn build, grouped and pointed at the milestone. If the raise came out of the model correctly, the allocation already sits inside it: headcount by function, the go-to-market spend that drives the revenue build, product and infrastructure, and G&A.

The test of a good breakdown is that every bucket maps to the milestone. It's the same discipline that governs the revenue side of the model - every projected unit tied to a driver - applied to spend: each allocation should answer "how does this get us to the thing that makes us fundable next?" A detailed, milestone-linked breakdown is itself a credibility signal. It's the difference between a number and a plan.

Why this closes the loop

Walk it back through the chain and the whole thing hangs together. The milestone defines what you must prove. The burn to reach it, plus a buffer sized to how long the next round now takes, gives you the raise. The dilution ceiling either confirms that number or tells you something has to change. And the use of funds is just that same burn, disaggregated and pointed at the goal. Every figure traces to the one before it.

That's what makes the ask defensible. You're not presenting a number and hoping it survives scrutiny, you're presenting the reasoning, and the number is where the reasoning lands. In a market where diligence has turned forensic and investors rebuild your model from scratch, a founder who can walk any figure back to a named assumption signals exactly what capital is looking for: someone who understands their own business well enough to be trusted with more of it.

A closing note on the figures in this article: they're 2025-2026 benchmarks from Carta, Y Combinator, a16z, and European VC data, and they'll drift with the market. Treat them as directional, and where a number is load-bearing for your own raise, check it against the latest primary source rather than any single guide, including this one.