Series A Benchmarks 2026: ARR, Growth, Round Size, Valuation

Series A benchmarks in 2026 come down to six numbers investors carry in their heads: what a round raises, what it's priced at, what it costs in equity, the ARR and growth that earn it, the efficiency behind that growth, and how long the whole thing takes. This page holds the data, with the period and the source on every figure. For the narrative, the KPI ladders by sector and the investor-relationship playbook, read our Series A fundraising guide; for the round before this one, see the seed round benchmarks.

In 2026 the median software Series A on Carta raises $14.4M at an $80M post-money and sells about a fifth of the company. The bar behind it is a few million dollars of ARR growing two to three times a year, with a burn multiple near or below one. In our work, the companies that clear it are the ones whose numbers tie out before the first meeting.

Read every row with its period. The Carta figures cover software companies on its platform; PitchBook-NVCA covers the whole US market; Crunchbase counts announced rounds worldwide. They don't agree, and that disagreement is the first lesson: there is no single Series A median, only the one for your sector, your geography and your quarter.

What is the typical Series A round size and valuation in 2026?

The median is now well into eight figures. Software Series A rounds on Carta raised $14.4M in the six months to July, and PitchBook-NVCA puts the median US Series A at $19.6M on a $62M pre-money. Both figures are lifted by AI companies: a non-AI founder should read the lower quartile and the non-AI split in the table, not the headline median.

Two things to take from the table. The median roughly doubled in a year on Carta's data, and the reason is the mix, not a kinder market: Peter Walker notes that at least half of the rounds behind the mid-2026 benchmark are probably AI-native (Carta). Strip the AI premium out and the non-AI Series A is priced far lower, which is why the quartiles and the non-AI split matter more than the median for most founders.

Europe prices the same company lower. Atomico measured the gap at roughly two-fifths on round size, with a similar discount on Series A pre-money, and PitchBook's European medians have been rising since (PitchBook). Our read, from raising on both sides of the Atlantic: the discount is a traction gap more than a geography tax. A European company growing at US rates gets close to US pricing; one growing at European rates does not.

How much dilution should you expect at Series A in 2026?

Median Series A dilution on Carta is 18%, the lowest in years, and founders typically keep a little over a third of the company once the round closes. Our healthy band is 18–25%, measured after the option pool refresh and the SAFE conversions, because a non-AI company priced below the median sells more of itself for the same cash. Model both before you talk to a lead.

Dilution is the one benchmark that hasn't moved much, and the one founders misread most. The headline percentage is what the new investor takes; the option pool top-up comes out of existing holders before the money lands, so real founder dilution runs several points higher. Our dilution guide walks the cap table round by round. The short version: quote post-pool, post-conversion ownership, because the associate on the other side will.

The band also depends on price. Carta's non-AI Series A median sits at a fraction of the AI figure (Carta, State of Private Markets), so the same median cheque buys a bigger slice of a non-AI company. That is why our band runs to 25%: it is an honest range for a company raising a real round at a fair price, not a target to negotiate towards.

What ARR and growth do investors actually expect at Series A in 2026?

For B2B SaaS, investors want a few million dollars of ARR growing two to three times a year. Point Nine's rule of thumb starts at $1M of ARR, and the seed investors Crunchbase quotes now cite $2M or more in the AI era. Below that, the round is judged on rate of change, not size; AI-native companies are held to a faster curve and priced on it.

The benchmarks disagree because they measure different populations. High Alpha's medians describe all private SaaS companies at a given ARR, most of which will never raise a Series A; Point Nine's napkin describes the ones that did. Read them together and the picture is consistent: the ARR floor has moved up since the 2021 market, and growth rate is what separates a fundable company from a healthy one.

The AI caveat matters. On High Alpha's 2025 survey, AI-native companies in the same ARR band grow at roughly three times the rate of classic B2B SaaS (Growth Unhinged). Investors price that curve, which is why the median Series A valuation looks unreachable to a non-AI founder and why the honest comparison set is your own category. Our Series A guide has the KPI ladders for consumer, marketplace, fintech, deep tech and hardware, where ARR isn't the metric at all; the SaaS metrics primer covers the definitions.

Which efficiency benchmarks do Series A investors check?

Three efficiency numbers get checked in every Series A data room: burn multiple (net burn divided by net new ARR, where under 2x is acceptable early and under 1x is excellent), net revenue retention (Bessemer's 'best' starts at 120%), and software gross margin, where private SaaS medians sit in the high seventies. We've seen more Series A processes stall on these three than on growth.

Burn multiple earns its place at the top because it is a catch-all: a gross margin problem, a churn problem or a sales-efficiency problem all show up in it eventually (David Sacks). The retention numbers earn theirs because they predict growth. In High Alpha's data, companies with high NRR and short CAC payback grow at more than double the rate of the rest (Growth Unhinged). An investor reading your KPI sheet is triangulating those two before they look at the top line.

Where founders lose the efficiency argument is the model, not the metrics. In roughly 60% of the models we review, first-year revenue projections are two to three times too aggressive, and an efficiency story built on that base collapses in the first diligence call. The fix is a driver-based forecast that ties to the KPI sheet and survives a stress test; our financial modeling team builds exactly that. Across the decks we've tracked, a dedicated financial-projections slide has correlated with about 40% more capital raised.

How long does it take to get from seed to Series A, and to close the round?

Two clocks. The gap from seed close to Series A close was 1.9 years at the median in late 2025, shorter than a year earlier but still beyond the old two-year plan. The process itself runs months, not weeks: founders raising solo typically take most of a year, and the raises we run close 70% faster on median because the materials and the target list exist before the first call.

The graduation numbers are the ones to sit with. In a normal year a quarter to a third of seed companies reach a Series A within two years, and the cohorts that raised at the 2022 peak did far worse (Carta). Crunchbase's recent cohorts look worse still, though part of that is simply time: companies that seeded in 2024 haven't had their two years yet. Either way, the base rate says most seed companies don't get there on the standard timeline, which is why the process has to start long before the process.

That is the argument for the relationship doctrine in our Series A guide: investors want to see a line, not a dot, and the active raise is the back half of a longer arc. Carta's own advice is to assume the money you raise now must last two and a half years (Carta). On the process itself, the phases we plan for are materials, outreach and first meetings, deep-dive diligence with the lead, then term sheet and close; that last phase alone runs six to eight weeks on Carta's estimate, and it is where an unindexed data room costs you the most.

What moved in 2026 versus 2024–2025?

Round sizes and valuations roughly doubled at the median, dilution stayed flat, and the time between rounds started to shorten for the first time since the reset. The money concentrated: AI companies took more than 60% of venture capital on Carta in Q1, and $100M+ Series A rounds are on track for a record. For a non-AI founder the ARR bar rose while the price paid for it did not.

The pattern is a barbell, not a boom. Megadeals took nearly nine-tenths of the US venture dollars deployed in the first half of the year (PitchBook-NVCA), and the same concentration shows up at Series A in the jumbo-round count. Medians rose because the top of the distribution pulled them; the company in the middle of a non-AI category is raising into a market that got choosier, not richer. Carta's down-round rate falling back to pre-pandemic levels is the good news in the table: terms are clean again, if you can get the round.

What these benchmarks mean for your raise

Use them to set the floor, not the ask. Benchmark against your category and quarter, price the round from a model that ties to your KPI sheet, and quote dilution after the pool and the SAFEs. In our work on 884 projects, the rounds that close fastest are the ones where the investor's rebuilt numbers match the founder's, so make the benchmarks yours before they become the investor's.

• Benchmark the right population. A vertical-SaaS company at $2M ARR compares to High Alpha's low-single-digit-millions band and Point Nine's napkin, not to Carta's AI-heavy median. Bring the comparison set to the meeting before the investor does.
• Let the model set the ask. The round size is the cash gap in a bottom-up forecast plus a buffer, not the median. Raise the median with no gap to fund and you've bought dilution you didn't need; our financial model work starts there, and the dilution guide shows what each extra point costs.
• Fix efficiency before growth. Burn multiple, NRR and CAC payback are adjustable within two quarters: move to annual contracts, cut the channels with the worst payback, expand the accounts you already have. Growth rate takes longer to move, and investors know which levers you pulled.
• Run the process like a sale. A short list of twenty to thirty thesis-matched funds beats a hundred names, and warm paths should cover most of it. A data room indexed to the lead's list shortens diligence by about 30%. Our due diligence checklist is the index; the investor databases guide is where the list starts.
• Keep the lines open. Monthly investor updates to the funds on your list turn the cold-start round into the back half of a relationship, and the data on time between rounds says you'll need that head start.

The 2026 Series A in one passage. On Carta's latest benchmarks, the median software Series A raises $14.4M. The post-money on that round is $80M. PitchBook-NVCA's whole-market median for the US is larger, at $19.6M, and both figures are lifted by AI companies. Point Nine's rule of thumb for SaaS is a few million of ARR growing two to three times a year. Bessemer's fundability bands put 'best' net revenue retention at 120% or more. David Sacks calls a burn multiple under 2x reasonable for an early-stage company. High Alpha's private-SaaS median gross margin sits in the high seventies. On timing, Carta's median gap from seed close to Series A close was 1.9 years at the end of last year. The median founding team owns about 36% once the round is done.

None of this requires a bigger number in the deck. It requires the right comparison set, a model the associate can rebuild, and a process that starts before the runway forces it. The raises we run through our Series A advisory begin with exactly that benchmark-and-model work, and it is why they close 70% faster on median than founders running alone. Series A is also the round we close most often, which is a different kind of benchmark.