The rule of thumb outlived the market that produced it

Ask a founder how much runway a seed round should buy and most will say eighteen months, maybe twenty-four if they have raised before. The number is not arbitrary. It comes from a market where it was roughly right: raise enough to hit the metrics that open a Series A, add a few months of slack for the raise itself, go.

J.P. Morgan examined companies that closed a Series A between the first quarter of 2025 and the first quarter of 2026, then worked backwards to look at the seed rounds those same companies had raised. The conclusion appears in the report without much hedging: relying on the eighteen to twenty-four month rule of thumb for fundraising "would have left some founders short."¹ That is a bank telling its own clients that the standard planning assumption has been drifting away from reality.

The drift is not news to anyone who has watched a friend spend nine months on a Series A process. What is less well understood is the shape of it. The story is not that everything moved back by six months in an orderly way. The distribution pulled apart, and the plans that break are the ones built around a single expected date.

The median moved, but the tails moved more

Carta's analysis of 3,365 US startups that went on to raise a Series A between the first quarter of 2018 and the third quarter of 2025 gives the clearest view of this. In the third quarter of 2019, 41 percent of companies raised their Series A within eighteen to twenty-four months of the seed round. By the third quarter of 2025 that share had fallen to 23 percent. Over the same period, the share taking three years or more went from 19 percent to 39 percent.²

Read those two pairs of numbers together, because separately each one understates the change. The band that the conventional planning assumption describes has shrunk by nearly half. The band that no conventional plan accounts for has roughly doubled. A founder who raises a seed round today and plans for the modal outcome of 2019 is planning for something that now happens to fewer than a quarter of companies. One caveat applies to the whole distribution: the sample is graduates only, companies that did go on to raise a Series A within the window Carta measured. It says nothing directly about the seed cohort that never reached one, which is the same selection problem the next section runs into from a different angle.

The medians tell a milder version of the same story, and they need a caveat. Carta put the median gap between a seed round and a Series A at about 2.2 years as of the end of 2024.³ For companies that closed a Series A specifically in the second quarter of 2025, the median interval was 616 days, a little over twenty months, which is shorter than the end-of-2024 figure and longer than the same measure two years earlier.⁴

That apparent reversal is a measurement artifact more than a market turn, and it is the kind of thing that gets quoted out of context. A quarterly median interval only counts companies that made it to a Series A in that quarter. It says nothing about the companies still waiting, and it moves whenever the mix of who graduates changes. In a market where more than 60 cents of every venture dollar on Carta went to AI companies in the first quarter of 2026,⁵ and where AI companies tend to raise faster, a falling median for graduates is consistent with a lengthening road for everyone else. The metric that describes your situation is the distribution across the whole cohort, not the median of the survivors.

Graduation rates point the same direction, with the same need for care about what is being counted. Carta's own figures put the 2018 seed cohort at something like 25 to 30 percent reaching a Series A within twenty-four months, against roughly 17 percent for the 2022 cohort.⁶ Those are twenty-four-month windows, not eventual outcomes, and the difference matters: a company that raises its Series A in month thirty-four has not failed, it has fallen outside the window the statistic measures. Any figure in this area should be checked for which window it uses before being compared to another one.

The median seed round is not the seed round that gets to Series A

The J.P. Morgan cohort work adds a detail that changes the practical advice, though not the one it is easiest to reach for. The report breaks down Series A deal size by the size of the seed round that preceded it, and that chart is about the Series A, not the seed: it shows that companies raising a larger Series A had typically raised a larger seed first. It is not a table of how long a given seed size takes to convert.

The more useful fact sits next to that chart. Companies that went on to successfully raise a Series A had typically raised a seed of four to seven million dollars.⁷ The all-seed median, across every company raising a seed regardless of what came next, has held between three and 3.3 million dollars since the start of 2025.⁷

Put those together and the median is doing exactly the misleading work the rest of this piece keeps finding. A founder raising the typical three-million-dollar seed round is not raising the seed size that typically converts. The companies that reach a Series A are disproportionately the ones that raised more at seed than the median company does, which is either a signal that better companies raise more, or that raising more helps you get there, or some mix of both. The data does not separate those, and neither claim licenses a specific number of months. What it licenses is narrower: the size of your seed round relative to the pool that actually converts is information worth having before you assume the median outcome applies to you.

The bridge round is now part of the plan, whether or not you planned it

Here is where the timeline problem shows up on a cap table rather than in a spreadsheet. In the second quarter of 2025, 16.6 percent of all venture capital raised on Carta came from bridge rounds, up from 11.8 percent a year earlier. At Series A the figure was 22.5 percent of cash raised.⁸

A bridge is not a failure. Plenty of them are the correct move, and in a market where an extension can carry a company to a materially better set of metrics, taking one beats raising a priced round on weak numbers. The problem is the ones that are not chosen so much as arrived at. A bridge negotiated with eleven months of runway, from a position of strength, prices differently from a bridge negotiated with four months left and no alternative. Existing investors know which situation they are in, and the terms reflect it.

The rise in bridge activity is, in aggregate, the market absorbing a timeline mismatch that individual plans failed to absorb. Companies sized their seed rounds for a road that turned out to be longer, and the difference got funded later, on worse terms than if it had been raised at the start.

What this changes about how you raise and how you spend

None of what follows is an argument to raise as much as possible.

Size the round against a milestone and a distribution rather than a date. The useful question is not how many months of runway the money buys. It is what metric opens your Series A, how long the plan needs to reach it, and what happens if that takes fifty percent longer than the plan says. If the answer to the last part is that the company dies, there is no slack in a plan operating in a market where four in ten companies take three years or more.

Then work out which lever you would pull, before you need it. A company with a credible path to extending runway by nine months without breaking the growth story has a different risk profile from one whose only option is a bridge. That path might be a slower hiring plan, a revenue line you deprioritized, or a smaller round now with a defined trigger for the next one. Naming it in advance is most of the value, because the same decision made at four months of runway is a much worse decision than the one made at fourteen.

It is also fair to ask an investor what gap they are expecting. The question runs diligence in the other direction, and the answer tells you something. An investor who says eighteen months in this market is either seeing something specific about your category or has not updated their priors. Either way you want to know before you sign.

When the longer road is not the binding constraint

This argument sits in tension with a case we have made before, that a valuation is a promise you have to keep and that raising the maximum available number is often the more expensive mistake. Both things are true, and they are not resolved by splitting the difference. Raising more money at a higher price makes the next round harder to justify on multiple grounds; raising too little for a road this long makes it harder to reach at all. The variable that reconciles them is not the size of the round but the price attached to it. A larger round at a valuation you can grow into over three years is a different instrument from the same amount of money at a valuation that assumed two.

There is a second qualification, and it matters more for some companies than the first. Averages describe a market, not a company. In a category where the market is being carved up quickly, or one where the technical build runs long before any revenue arrives, undercapitalizing remains the more common way to die, and a founder in that position should weight the timeline data accordingly rather than treat it as a reason to raise less. The same goes for the fastest-moving AI companies, where the constraint is rarely time to the next round. If you are in one of those situations, the useful takeaway from all of this is narrower: know that you are the exception, and know why, rather than assuming it.

For most founders raising a median-sized seed round, though, the arithmetic is not complicated. The road has lengthened, the distribution has widened, and the eighteen-month figure most plans still use was calibrated for a market that ended several years ago. Changing the number you plan against costs nothing at the point where you are still modelling the round. It costs a great deal at month twenty.

The harder part is working out where your company sits in that distribution, which depends on your category, your seed size, and what the investors you are targeting expect. That is a question about your position relative to a pipeline of comparable companies, which is the kind of thing AngelHive's assessment is built to answer.

Sources

1. J.P. Morgan, "Startup Insights Report for the Innovation Economy," H1 2026 edition, published 7 May 2026. https://www.jpmorgan.com/content/dam/jpmorgan/documents/cb/insights/banking/commercial-banking/cb-insights-banking-startup-insights-report-h1-2026.pdf

2. Carta data on 3,365 US startups that raised a Series A between Q1 2018 and Q3 2025, as reported and analyzed by SaaStr. https://www.saastr.com/your-seed-round-now-needs-to-last-3-years-what-3365-startups-tell-us-about-the-new-series-a-timeline

3. Carta, median time between rounds as of end of 2024, as reported by SaaStr. https://www.saastr.com/carta-the-average-time-from-seed-to-series-a-has-hit-2-2-years-and-longer-from-series-a-to-series-b

4. Carta, "Series A funding slides in Q2 2025." https://carta.com/data/series-a-fundraising-q2-2025/

5. Carta, "State of Private Markets: Q1 2026." https://carta.com/data/state-of-private-markets-q1-2026/

6. Carta, "Graduation rate from seed to Series A." https://carta.com/data/newsletter-graduation-rate-from-seed-to-series-a/

7. J.P. Morgan H1 2026 Startup Insights Report, median seed deal size, corroborated by PitchBook-NVCA Q1 2026 Venture Monitor. https://www.jpmorgan.com/content/dam/jpmorgan/documents/cb/insights/banking/commercial-banking/cb-insights-banking-startup-insights-report-h1-2026.pdf

8. Carta, "Bridge rounds got a boost in Q2." https://carta.com/data/bridge-rounds-q2-2025/

Runway

The Road to Series A Is Longer Than Your Runway Plan