The companion study to The Series B Clock. Same 615,718 SEC Form D filings, same survival model, one rung higher. A Series B company starts with better odds than a Series A company — and then loses them faster.
Series B companies start at 35.5% odds versus 30.3% at Series A, but their hazard spikes harder, decays sooner, and by month 42 a stalled Series B company is in worse shape than a stalled Series A company.
Series B vintages 2014–2019 (n = 645), each observed for 78+ months. Median detected Series B is $25.9M and median graduating Series C is $55.0M — both squarely on published medians.
01
Five points better than the 30.3% we measured from Series A. Survivorship is real — a company that cleared the B gate is a better company. The range across definitions is 24–44%.
02
Both jumps peak in the quarter ending month 21. But the Series C hazard peaks at 5.5% per quarter against 3.7% for Series B — a spike half again as tall, and half again as narrow.
03
B→C conditional odds start higher and decay faster. They cross A→B at month 42 and stay below it. A stalled Series B is a harder problem than a stalled Series A, because the burn base and the valuation to clear are both far larger.
04
Between $15M and $45M, round size barely matters — all three bands land at 34–36% by five years. Above $45M it falls off a shelf to 24.6%. A big B buys a harder exam, not more time.
05
Of the 200 companies that closed a Series B in 2022, zero raised a Series C within twelve months. Five made it by month 18. Twenty-one have made it at all.
06
Of those who never raise a C, 63% never file another financing of any size — the same disappearing act we saw one rung down, at a much higher burn rate.
So what
Clearing the Series B gate buys you a better starting position and less time to use it. The window that matters is months 12–24 after the B closes; past month 33 the hazard has already halved, and past month 42 you are statistically worse off than a company a full stage behind you.
Conditional probability that a company still without a Series C at month m will ever raise one. Kaplan–Meier estimate, Series B vintages 2014–2019 (n = 645).
Both rungs measured identically, on the same filings, with the same model.
Probability a still-waiting company ever raises the next round · the lines cross at month 42
Series B companies begin five points ahead and end three points behind. The crossing point is the practical definition of "stalled at Series B."
Share of still-waiting companies closing the next round each quarter
Same peak month, very different peak height. The Series C spike is 48% taller and falls below half-peak three months sooner.
Only 4% of the cohort has raised. The scaled go-to-market machine that gets diligenced next year is being assembled right now.
Half of everyone who will ever raise a C has done it. Series B runway is spent, and the burn base is three to four times what it was at Series A.
The hazard drops below half its peak — three months earlier than the equivalent moment at Series A. This is where the Series C window actually shuts.
Odds fall below those of a stalled Series A company. From here the realistic outcomes are profitability, a sale, or a structured round.
Conditional odds of ever graduating, by how strictly a Series C is defined
The level moves with the definition; the decay does not. Under all three, roughly four-fifths of the day-one odds are gone by month 48.
Share of still-waiting companies that close a Series C each quarter · Series B vintages 2014–2019
The window opens at month 9, peaks in the quarter ending month 21, and falls below half its peak at month 33.
Cumulative graduation (purple) against conditional probability of ever graduating (lime)
Cumulative share of each Series B vintage that had raised a Series C by 12, 24 and 36 months.
Kaplan–Meier cumulative incidence · bars omitted where the vintage is not yet old enough to observe
The 2019 vintage is the strongest on record at 43.0% by five years. The 2022 vintage recorded no Series C rounds at all in its first twelve months.
A caution
Among the few 2021–2023 vintage companies that did reach a C within three years, they got there in a median of 13.2 months versus about 20 months for older vintages — but that is only 69 companies. Read it as bifurcation, not acceleration: in that window a small set of mostly AI-adjacent companies were funded very fast while everyone else was funded not at all.
Six in ten Series B companies raise something again within five years. Only three and a half in ten raise something bigger. The gap is the same bridge economy we found one rung down — but the cheques are larger and the dilution is worse.
Of the 414 companies in our mature cohorts that never reached a Series C, 261 (63%) never filed another financing of any kind.
As at Series A, an interim round correlates with better outcomes, not worse: among companies still C-less at month 24, those who took a smaller round went on to a real C 30.7% of the time versus 19.1% for those who did not. Insider conviction remains the best observable signal on the board.
Cumulative share of the 2014–2019 Series B cohort
Share raising a Series C within 60 months · 2014–2019 vintages
At Series A the relationship was an inverted U — too small hurt, too big hurt. Here it is a shelf. From $15M to $45M the five-year odds sit in a flat band of 33.7–35.9%, differences well inside the noise. Then the $45–75M band drops to 24.6%.
A $60M Series B implies a post-money in the $250–350M range. The Series C that clears it has to be priced above that, on metrics that have to have roughly tripled. The round that felt like validation is the round that sets the bar you then fail to clear.
At the three-year mark the decline looks more gradual — 29.8%, 27.6%, 26.9%, 18.5% — which says the mid-sized rounds are not less likely to get there, just slower. Only the largest band never catches up.
Geography matters less here than at Series A: 36.1% in NY/MA and 34.0% in California against 31.6% elsewhere — a four-point spread, down from six at the earlier stage.
The $45–75M band is 65 companies; treat that bar as directional rather than precise.
The arithmetic is almost identical to the Series A study — but what has to be true at the end of it is materially harder.
Series C closes. Median for companies that graduate within three years.
Process starts. A Series C takes 4–6 months, and later-stage diligence is heavier — cohort data, capacity models, quality of revenue.
The diligence window. Growth-stage investors want three to four consecutive quarters of a machine, not a motion: quota attainment across a full team, payback holding as spend rises, net revenue retention at scale.
The machine has to be running. Segmented pipeline, a capacity model that ties heads to bookings, clean cohort retention. At Series A you are proving a motion works; at Series B you are proving it survives being multiplied.
Series B closes. The instrumentation has to be built before the hiring starts — you cannot reconstruct clean cohort and capacity data after the fact.
A Series A is won by proving one motion repeats. A Series C is won by proving the motion survives being multiplied by ten. Both are decided in the first nine months after the money lands.
Source
Every quarterly SEC Form D structured data set from 2014 Q1 to 2026 Q2, collapsed into 32,717 financing events across 18,997 US technology companies. Identical pipeline to the Series B study.
Definition
A "Series B" is a company's first equity event of $15–75M that is a ≥1.3× step-up on its prior maximum and is not its first financing. A "Series C" is a later event of at least max($25M, 1.3× the B).
Validation
Median detected Series B is $25.9M and median graduating Series C is $55.0M, against published medians of roughly $27–38M and $50–60M. Median time B→C among graduates is 23.6 months.
Limit 1
645 mature Series B companies against 3,093 at Series A. The headline curve is solid; the sub-cuts by round size and geography are directional. Cells under 40 companies are suppressed.
Limit 2
Some rounds never file, or file under a different classification. This biases the graduation level down — a widely repeated industry figure puts Series B→C nearer 60% — but it does not bias the timing, which is the finding.
Limit 3
This cohort is, by construction, companies good enough to have raised a Series B. The 35.5% is not comparable to a population rate, and the comparison to Series A is a comparison of conditional odds at each rung, not of company quality.
To go deeper
The same three gaps as the Series A study, and one more that bites harder here: (1) real round labels and sector tags — at Series B the size bands overlap between stages, so labelled data would sharpen the cohort materially; (2) outcomes — at this stage many non-graduates exit rather than die, and Form D cannot see an acquisition; (3) performance — ARR, growth and net retention, which is what actually separates the 35% from the 65%.