The 2026 SaaS Funding Guide

The 2026 SaaS Funding Guide

By Ryan Allis, CEO of SaasRise · August 2026

Venture capital came back in 2026, but not for everyone. Carta counted $30.4 billion in the first quarter and down rounds fell to 11.4%, while 83% of software dollars went to AI companies. This report takes the bar apart stage by stage: the ARR, growth rate and retention a seed, a Series A and a Series B actually require in 2026, what each round costs you in ownership, where the gross margin bar now sits for SaaS versus AI, and what sixteen companies that closed rounds this year had on the board when they did.

The bar at a glance

Everything in this report expands on the rows below. They are the numbers a 2026 round is priced against for a B2B SaaS company that is not a foundation-model business. Sections 2 to 4 explain where each figure comes from and how much room there is around it.

Ranges are for B2B SaaS raising in 2026, drawn from Carta round benchmarks, Crunchbase seed and Series A data, SaaS Capital and Benchmarkit. AI-native companies are priced against a different set of comparables, covered in sections 1 and 5.

Key findings

The bar rose at every stage, but it rose unevenly and for different reasons. At seed the change is that revenue is now expected where a prototype used to be enough. At Series A the change is scarcity: fewer companies graduate, they take longer to do it, and the ones that clear are priced against an AI comparison set that has nothing to do with them. At Series B the change is that growth alone stopped being sufficient. The round is underwritten on whether a dollar of capital reliably becomes a dollar of durable recurring revenue.

The one number that changed most between last year's playbook and this one is gross margin. It used to be a single expectation: 75% or better, or explain yourself. In 2026 it is two separate bars: traditional SaaS software revenue still runs at an 80% median, while the average AI product runs at 53%. Section 5 covers how investors are now underwriting that gap, because getting it wrong in either direction costs you the round.

Three things are worth internalising before you read the stage sections. Median seed post-money on Carta rose from $15.0M in 2021 to $21.6M by Q1 2026, which means the price you set at seed is now a much higher hurdle to grow past. Only 17% of the 2022 seed cohort reached a Series A inside two years, against 25–30% in a normal vintage. And a non-AI startup raising a Series A in Q1 2026 was priced at a $55M median while an AI foundational-model company at the same stage was at $300M. You are not competing with that company for capital, but your board will still see its valuation on a slide.

1. The market you are raising into

Carta's Q1 2026 data describes a market that has fully cleared the 2022 hangover. Down rounds fell to 11.4%, back in line with 2019 and 2020. Median dilution across seed through Series C dropped from about 18% to 16% over the year, and at Series B it fell to 12.9%. Terms favour founders again. Series B and Series C pre-money valuations rose 17.2% and 12.5% respectively year over year.

Median software rounds over the past six months. Dilution: 18% at seed, 18% at Series A, 12% at Series B.

Those are good numbers. The problem is what sits underneath them. The PitchBook-NVCA Venture Monitor for Q2 2026 puts the concentration in plain terms: AI accounted for 86% of all venture dollars in the quarter, and three firms (Andreessen Horowitz, Thrive Capital and Founders Fund) took in 48.1% of all capital raised by venture funds. First-time fund formation is on pace for its lowest year since 2016. The market is setting records at the very top and contracting almost everywhere underneath it.

Total round count in 2025 fell to a six-year low, and the same pattern held into 2026: more money into fewer companies. In pre-seed, U.S. startups on Carta raised $3.19 billion across roughly 11,500 instruments in Q2 2026, against $3.22 billion across 14,825 instruments a year earlier. Nearly identical dollars, 22% fewer deals. AI took half of pre-seed dollars.

For an operator this has a specific consequence. Median statistics now describe a barbell rather than a middle. When Carta reports a $80M median Series A post-money for software, that median is pulled by a cohort of AI companies raising at prices your business will never see, and it sits alongside a long tail of $25M–$50M pre-money rounds that look nothing like the headline. Benchmark yourself against the non-AI figure, $55M median at Series A in Q1 2026, and treat the blended median as market colour, not as a target.

Two companies at the same stage, priced five and a half times apart.

2. Seed: the bar is now revenue, not a deck

The clean line between pre-seed and seed in 2026 is whether an investor needs to look at a customer list to say yes. Pre-seed still buys a team, a thesis and a prototype, typically $500K to $2M on a post-money SAFE with a cap around $10M–$15M. Seed buys evidence. The median U.S. seed round is roughly $3M by Crunchbase's count and $4.1M in Carta's software-only sample, priced at a $24.3M median post-money, with 18% dilution.

What has actually changed is the revenue expectation. Andy McLoughlin, managing partner at Uncork Capital, told Crunchbase in May 2026 that his firm's typical seed check has almost doubled in 18 months, from $2.5M or less to $4.5M, while the firm still targets at least 10% ownership. Bigger checks at higher prices mean more proof required. The working expectation for institutional B2B software seed rounds now sits around $300K–$500K of ARR with ten to twenty paying customers, where $200K would have carried the day two years ago.

That is the median case. The rounds that got written about in 2026 cleared it by a wide margin.

Every figure here comes from the company's own announcement or first-party press coverage. None are audited.

Notice what the investors quoted in these rounds actually talked about. Alexander Kölpin at seed + speed said Promptwatch convinced them because the team "achieved more than 2 million euros in ARR with limited capital and personnel resources." Blum Ventures' André Hammerer made the same point: €2M in twelve months on €1.2M and a small team. Neither mentioned the model, the market size or the roadmap. They priced the ratio of output to input.

What a seed investor is underwriting

Strip away the deck advice and there are four things a seed lead is checking in 2026, in this order:

• Retention with a curve you can show. Eight or more weeks of stable or rising cohort retention is the hardest signal to fake, which is exactly why it carries the most weight. Cumulative signups carry none.
• Money that actually changed hands. A $2K paid pilot from a named buyer outranks a $200K letter of intent. LOIs without a specified scope and date get discounted to zero.
• Capital efficiency. The Promptwatch pattern. At seed, investors want a burn multiple comfortably under 1.5x, meaning less than $1.50 of cash for each $1 of new ARR.
• A named wedge, not a category. A proprietary dataset that compounds, a distribution advantage, an integration nobody else has. Feature-level differentiation does not survive diligence in a market where competitors ship weekly with AI assistance.

3. Series A: the hardest round in venture

The Series A is where the 2026 market breaks most founders' plans. Not because the money disappeared. Median Series A round size in Carta's software sample is $14.4M at an $80M post-money, and Crunchbase puts the all-sector median at $15M with an upper quartile of $25M. The money is there. The problem is who gets to see it.

Graduation within 24 months, by seed vintage. The 2022 cohort had the metrics bar move underneath it mid-flight.

Carta's own analysis is blunt about the cause: "The metrics needed to raise a Series A shifted underneath many of these founders and they've struggled to keep up with the new, higher requirements." The picture worsened steadily and then began to turn. Among startups raising a Series A in Q4 2024, the median interval since the seed round was 774 days, about 2.1 years and 84% longer than the 420 days a company raising in Q4 2021 had taken. By Q4 2025, across 9,843 rounds, Carta had the median seed-to-A gap back down to 1.9 years. It is improving, but it is still five to eight months longer than the eighteen-month runway most seed rounds are sized for.

That arithmetic is why seed extensions stopped being a distress signal. A $1.5M–$3M top-up on a flat or slightly stepped-up SAFE is now a normal planning instrument, because the distance between what a seed round bought and what a Series A demands takes longer than one seed round to cover. If you need one, price it as a deliberate step rather than a rescue, and be specific with your existing investors about the milestone it buys.

What ARR do you actually need?

There is no audited survey publishing a median ARR at Series A close. The tables circulating online that claim one trace back to content-marketing sites recycling each other. What does exist is a consistent practitioner range and a set of disclosed rounds, and between them they give you a usable answer.

McLoughlin's version, on the record: "The threshold for raising a successful Series A is no longer $1 million in annual recurring revenue. In the AI era, startups are expected to show $2 million to $3 million — even $4 million — in ARR as proof that the business has the momentum to scale." That matches what founders report from processes in the first half of 2026: roughly $1M is the floor to get meetings if the growth curve is loud, $2M–$3M is where a competitive process happens, and below $1M you are raising a seed extension whatever you call it.

What the 2026 Series A announcements had on the board

Two patterns run through that table. First, almost none of these announcements led with an ARR number. They led with a growth multiple. Netris said 800%, Prelude said 6x, Caruso said 400%, ProLine said 4x. Companies disclose the metric that flatters them, and in 2026 the flattering metric is the slope. Second, three of the seven are not AI-model companies at all. ProLine sells to roofing contractors and Caruso does fund admin in Australia and New Zealand. Vertical software solving a mission-critical workflow inside a traditional industry is still a fundable Series A, and it clears at $9M–$12M round sizes rather than $38M.

Round size tracks the growth number in the press release far more closely than it tracks revenue scale.

4. Series B: proof that the machine works without you

Series B pricing improved in 2026. Pre-money valuations rose 17.2% year over year, the median software Series B in Carta's sample closed at a $191M post-money on $25M raised, and dilution at the stage fell to 12%. Founders are selling less of the company for more money than at any point in the last few years.

What they are selling has changed. A Series A is underwritten on a growth curve. A Series B is underwritten on a system: whether a dollar of sales and marketing spend reliably produces a predictable dollar of recurring revenue when the founder is not in the room. That shift shows up in which metrics kill deals. Growth below 60% year over year at this scale makes a competitive process very hard to run. NRR below 110% invites questions the topline cannot answer. A burn multiple above 2.0x, more than $2 of cash burned per $1 of new ARR, reads as a company that has bought its growth rather than built it.

The disclosed 2026 Series B rounds sit at very different revenue scales, which is what a stage looks like when $10M and $60M of ARR can both clear.

Omilia and Respond.io are the two worth studying if you run a private SaaS company outside the Bay Area. Omilia grew from roughly $6M to over $60M of ARR on a $20M round raised in 2020 and nothing since. Respond.io reached $35M ARR at a 30% profit margin before taking growth capital, on a $7M Series A from 2022. Both went to market from a position where the round was optional. That is what produced the pricing, not the AI label.

5. Gross margin: the SaaS bar and the AI bar are now different numbers

For fifteen years gross margin was the least interesting slide in a SaaS deck. You showed 75% or better, the investor nodded, everyone moved on. That single expectation has split in two, and mixing the two up is now one of the more expensive mistakes a founder can make in a diligence process.

Start with traditional software, because the news there is better than the noise suggests. Benchmarkit's 2026 SaaS and AI-Native Metrics study, covering 342 companies on full-year 2025 data, puts median software gross margin at 80%, flat across four years. Blended across total revenue, including services, the median is 76%. The top quartile clears 86%. Industry-wide AI infrastructure costs have not compressed software margins at the median, and any investor telling you otherwise about a subscription business is arguing with the data.

Two exceptions inside that dataset matter. Companies on usage-only pricing run at 62%, because compute and infrastructure scale directly with consumption. And the bottom quartile of all software companies sits at 50%, which is where per-customer customisation, heavy services and unmanaged infrastructure end up.

Five different businesses that all get described as software. Investors no longer benchmark them against each other.

The AI numbers, and why they are rising

ICONIQ surveys roughly 300 software executives, mostly at companies below $100M of revenue, for its State of AI research. Its picture of AI product margins is a clear upward line: 41% in 2024, 45% in 2025, a projected 53% in 2026, and 59% forecast for 2027. That is fourteen points of expansion in two years, and ICONIQ attributes it to falling inference costs, model routing and caching strategies, and revenue scale, not to pricing power. The supporting detail is convincing: 66% of respondents improved their cost per query by 10% or more, and 19% improved it by 30% or more.

The margin panic of 2024 is over, but the gap to traditional SaaS is still more than twenty points.

The layer you operate at changes the number materially. Application-layer AI products ran at 33% in 2024 and 38% in 2025, with roughly 45% projected for 2026 and 60% by 2027. Infrastructure and platform AI products are projected to reach 67% by 2027, higher than the application layer, which inverts the usual assumption that infrastructure is the capital-heavy, low-margin end of the stack. High-growth companies are projected at 64% in 2027 against 58% for everyone else, so the margin and the growth rate move together rather than trading off.

At the extreme end, Bessemer's State of AI research found that the ten fastest-scaling AI startups it surveyed — companies reaching roughly $40M of ARR in their first commercial year and $125M in their second — averaged about 25% gross margins, sometimes negative, deliberately buying distribution. The same cohort produced $1.13M of ARR per employee, four to five times a typical SaaS benchmark. Bessemer's slower, more capital-efficient group, which it calls Shooting Stars, reached around $3M of ARR in year one at roughly 60% gross margins. Two completely different financial profiles, both raising at premium valuations.

What to actually target, by what you sell

Three things follow for a fundraise. First, segment the number before an investor does: show software margin, AI-attributable COGS and services margin as separate lines, because a blended 62% reads as a broken SaaS business and reads as a healthy AI business, and the difference is entirely in the presentation of facts you already have. Second, show the trend and the mechanism. A margin that went from 38% to 52% because you moved 70% of calls to a smaller routed model is a far stronger slide than a static 60%. Third, name the floor. Most Series B investors underwrite a path to 70%+ for anything sold as software. If your architecture cannot get there, you need to be raising against the AI benchmark set and saying so explicitly.

6. Growth, efficiency and the metrics that decide the round

SaaS Capital's 14th annual survey of more than 1,000 private B2B SaaS companies put the 2025 median growth rate at 22%, down from 25% the year before and 30% in 2023. Only 7.3% of companies reported flat or negative growth, so this is not a market falling apart. It is a market growing steadily and slowly. Bootstrapped companies between $3M and $20M of ARR grew at a 15% median, with net revenue retention of 103% and gross revenue retention of 91%.

Now compare that with the growth rates venture investors are pricing.

ICONIQ's top-quartile companies under $50M ARR grew 111% in H2 2025. The median private B2B SaaS company grew 22%.

The gap between those two lines is the whole story of the 2026 funding bar. Venture-priced rounds are underwritten against the top quartile of a top-decile investor's portfolio, not against the market. A company growing 40% a year at $8M of ARR is performing above the median of every private SaaS survey published this year and is still, in venture terms, an awkward fundraise. That is not a judgement about the business. It is a statement about which pool of capital fits it.

The four numbers that get argued about in the partner meeting

Retention, benchmarked by contract value rather than by stage. SaaS Capital's retention brief makes the useful point that companies sharing a price point have more in common than companies sharing an age or an ARR band. Median gross revenue retention is 95% for companies with average contract values above $250K and roughly 91% below that. Benchmark yourself against your own ACV band; a 91% GRR on a $15K ACV product is normal, and the same number on a $400K enterprise contract is a problem.

Burn multiple, where the distribution is stranger than the median. Lighter Capital's analysis of its private B2B SaaS portfolio found a median burn multiple of 0.79x, but among cash-burning companies the distribution is bimodal: about half burn more than $1 for every dollar of new ARR, about a quarter burn less than $0.33, and almost nobody sits in the middle. There is no gentle slope here. Companies are either efficient or they are not, and investors have learned to read the number as a binary.

CAC payback, on a gross-margin basis and by segment. Before the numbers, the definition, because half the arguments about this metric are really arguments about arithmetic. Every benchmark quoted below is gross-margin adjusted — months of gross profit, not months of revenue. The standard formula, as defined by the SaaS Metrics Standards Board and used by Benchmarkit, is sales and marketing spend for the prior period divided by (new ARR added × gross margin %), expressed in months. The customer-level version is the same idea: CAC ÷ (monthly ARPU × gross margin %).

That distinction is not academic. At an 80% gross margin, dividing by revenue instead of gross profit makes your payback look about 20% better than it is. At a 50% AI-native margin it makes it look twice as good. Two companies with identical CAC and identical ARPU can report 12 months and 19 months purely because one is a subscription business and the other is running inference. That is why the margin section above changes your payback number without anything in your sales motion changing at all. If you show the revenue-basis version in a deck, label it, because an investor will recalculate it on gross profit and the correction is not a good moment.

On that basis: Benchmarkit's 2026 study, using full-year 2025 actuals from 342 SaaS and AI-native companies, puts the median CAC payback at 16 months, with the first quartile at 10 months and the fourth quartile at 24 months or more. Optifai's larger sample of 939 B2B SaaS companies, using the customer-level formula, puts the median at 15 months and breaks it out by segment: 8–12 months for SMB products under a $15K ACV, 14–18 months for mid-market, and 18–24 months for enterprise deals above $100K. Under 12 months is best in class; above 24 months is a conversation you will not win. Benchmarkit publishes its methodology and Optifai is a vendor dataset, so use Benchmarkit for the level and Optifai for the segment shape.

The Rule of 40, and how companies actually clear it. Standard Metrics ran the framework across 1,377 venture-backed private companies with over $1M of annualised revenue and found that when private companies clear the Rule of 40, 89% of them do it through growth alone. Only 11% clear it through the balance of growth and margin the rule was designed to reward. If you are planning to make the number by cutting your way to profitability at a 15% growth rate, you will be in a category of one in the room, and not in the flattering sense.

Revenue-based financing, for the companies venture is not built for

If you do not want to sell equity, or your growth rate will not clear a venture bar, revenue-based financing is the instrument that fits the gap. You borrow a fixed amount and repay it as a percentage of monthly revenue until you hit an agreed cap, so payments flex down in a slow month and up in a fast one, and nobody takes a board seat or a liquidation preference. Lighter Capital, one of the longer-running lenders in the category, describes the standard shape: facilities up to roughly $4M, repayment as a set share of monthly revenue, terms of up to three years, and a cap usually between 1.35x and 2x the principal. That cap is the whole cost of the capital, which makes the arithmetic unusually easy to compare against the alternative. SaaS Capital, which both lends and publishes research on the question, puts the cost of equity for a private SaaS company well above the cost of debt, and notes that a company above about $3M of ARR with predictable retention can usually access a credit facility priced below a revenue-based loan.

The trade-off is real: the money comes out of cash flow from month one, so it funds things that pay back inside the term, sales hires into a working motion, a proven ad channel, an acquisition of a small competitor. It does not fund a two-year platform rebuild. Used the wrong way it is expensive money for a company that cannot service it. Used the right way, on a business growing 20% to 40% with 90%+ gross retention, it buys the growth without resetting the cap table, and it leaves the equity round available later at a price your own progress has set.

7. What each round costs you in ownership

Dilution is the quietest number in a term sheet and the one founders regret most. Carta's Founder Ownership Report puts median founding-team ownership at 56% after a seed round and 36% after a Series A, based on rounds raised from 2021 through 2025. By Series B, AI founding teams hold a 27.3% median stake and non-AI teams hold 21.8%. At Series C the median employee option pool (16.8%) overtakes median founder ownership (16.1%). The sector split shows up earlier than most founders expect: at Series A, founders in digital industries retain 37.5% while founders in physical industries retain 30.5%.

Founder ownership through the stages. The AI premium buys founders roughly five and a half points more of their own company by Series B.

The good news is that per-round dilution keeps falling: 18% at seed and Series A, 12% at Series B in Carta's recent software sample, against roughly 19% across those stages two years ago. Put a dollar figure on it. Sell 18% at an $80M post-money Series A and you have given up $14.4M of value at today's price. If the company is worth $500M at exit, that same 18% is $90M. Every point of dilution you avoid at the A is worth roughly $5M of exit value in that scenario, which is why the cheapest capital is almost always the revenue you generate before you raise.

8. If the goal is an exit rather than a round

Not every company on this path is raising a Series B. For plenty of SaasRise members the real question is what the business is worth to an acquirer, and the answer in 2026 is more sobering than the venture headlines suggest. Aventis Advisors put the median public SaaS EV/Revenue multiple at 3.4x as of March 2026, a significant decline as public investors discount software on AI disruption fears. Across 543 private SaaS M&A transactions with disclosed multiples since 2015, the median is 4.5x revenue, with a first quartile of 2.4x and a third quartile of 8.1x on a median deal size of $80M.

That spread — 2.4x to 8.1x — is where the work is. The same revenue base sells for three times more or three times less depending on growth rate, retention, gross margin and how concentrated the customer list is. Every metric in this report that gets you a better round also gets you a better multiple, and the two processes reward the same twelve months of preparation. The difference is that an acquirer will also look hard at whether your growth survives without the current spend, which is an argument for building acquisition channels that keep producing after the diligence closes.

9. The bear case

Everything above describes a market that works if you are on the right side of it. Four things could make this report look optimistic within a year.

The concentration is extreme and it is new. AI took 86% of venture dollars in Q2 2026 by PitchBook's count, and three firms raised 48.1% of all new fund capital. If AI valuations reset, the money does not automatically redistribute to non-AI SaaS — funds mark down, LPs slow commitments, and the whole early-stage market gets quieter. The 2022 seed cohort is the case study for what happens when the bar moves after you have already raised.

Enterprise AI revenue has not proved it is durable. Madrona surveyed 150 senior enterprise decision-makers in August 2026. Three quarters plan to expand AI budgets over the next twelve months, and 83% converted fewer than half of their AI pilots into production over the last twelve months. More than a third converted fewer than one in four. Just 1% got past three quarters. A lot of the ARR being priced at a premium right now is pilot revenue that has not yet survived a renewal cycle.

Fewer deals is the signal, not more dollars. Round counts fell to a six-year low in 2025 and pre-seed deal counts dropped 22% year over year while dollars stayed flat. Rising medians in a market with falling deal counts partly reflect survivorship: the companies getting priced are the ones that could get priced.

The valuations set in 2025 and 2026 have to be grown into. A $24M median seed post-money means the next round needs to clear roughly $50M to feel like a step up. Down rounds are at 11.4% today, but today's rate reflects companies raised at 2022 and 2023 prices. The cohort priced at 2026 highs has not faced its next round yet.

None of this argues against raising. It argues for raising an amount you can grow into, and for treating the milestone rather than the money as the point of the round.

10. Running the round once the numbers are there

Everything above is about being fundable. This section is about the ten to sixteen weeks in which being fundable turns into money in the bank, which is a separate skill and the one most first-time raisers underrate. A company with a Series A metric profile and a badly run process routinely loses to a weaker company with a well run one, because a fundraise is priced on competition and competition is something you manufacture.

Start when the metric arrives, not when the cash runs low. Work the arithmetic backwards. A process takes three to four months from first meeting to wired funds, diligence and documents included. You want to be negotiating with at least six months of runway behind you, because the moment an investor can calculate that you have ten weeks of cash left, your valuation stops being a function of your growth rate. That means the trigger to begin is a milestone on the board, not a number in the bank account. If you are four months from the ARR that makes the round obvious, wait the four months and raise a bridge if you have to. If you are fourteen months away, you are not preparing a Series A, you are preparing an extension, and pretending otherwise burns your best investor relationships on a pitch they will remember when you come back.

The funnel is much wider than founders expect. Y Combinator tracked its own companies through the Series A process and published the conversion rates. Companies that raised took an average of 30 coffee meetings with individual investors. Half of those produced a pitch to an individual partner, about 30% of partner pitches got to a full partnership meeting, and one in five partnership pitches produced a term sheet. That is roughly thirty first conversations for one offer. The median company in that dataset heard no 18 times before hearing yes, and one company was rejected 30 times before getting a term sheet it was happy with.

The Series A funnel measured across Y Combinator companies. The numbers date from 2019 and the shape has not improved since.

Two things follow. First, build a target list of 60 to 80 firms, not 15, and qualify them properly: check that the fund invests at your stage, in your category, at your cheque size, and that it has not already backed a direct competitor. Second, do not read early rejections as a verdict on the company. In a funnel with those conversion rates, a run of five no's is the expected outcome, not a signal to reprice.

Run it in parallel or do not run it. This is the single most important mechanical decision in a fundraise. YC's guidance to its own founders is to compress each stage into a window: all first pitches inside one to two weeks, all partnership pitches inside a separate one to two weeks, so that offers arrive at roughly the same time. Sequential processes, where you talk to one firm, wait, then talk to the next, produce exactly one term sheet at a time and hand the timing advantage to the investor. Parallel processes produce simultaneous offers, and simultaneous offers are the only reliable source of negotiating leverage a founder has. Practically: do not take a single meeting until the list, the deck, the model and the data room are all finished, because the first meeting starts a clock that runs on the investor's calendar, not yours.

Diligence now happens before the term sheet, not after. This is the change that catches founders who last raised in 2021. Customer reference calls, cohort retention analysis and competitive work that used to sit in the four weeks between signing and closing have moved forward into the decision process. Madrona, writing to founders in January 2026, describes what firms now expect to see up front: clean metrics, clear momentum, a coherent wedge, and customers who would genuinely struggle without you. The practical implication is that your data room needs to be complete on day one of the process, not assembled in week nine. Monthly P&L, balance sheet and cash flow for 24 months, a cohort file that ties to your ARR number, the cap table, customer contracts, and a reference list of five to eight customers who have agreed in advance to take a call. Metrics that shift between your deck and your data room are the most common way a warm process goes cold.

Know what is actually negotiable. Cooley publishes deal terms from the venture financings it handles, which makes it one of the few running counts of what is market rather than what is rumoured. In Q2 2026, across 166 financings, 95.8% of deals carried a 1x liquidation preference and 96.4% used nonparticipating preferred. Redemption rights appeared in 5.4% of deals and accruing dividends in 3%. Structure, in other words, is rare. If you are offered a 2x preference or participating preferred, that is not the market, it is a statement about your company, and the right response is to ask what the investor is worried about rather than to accept it as standard. What is worth your energy instead: the option pool, because a pool created pre-money is dilution paid entirely by you and existing holders, and board composition, because it outlives the valuation. Pro rata rights and standard four-year vesting with a one-year cliff are not worth burning credibility over.

Expect the price to move underneath you. In Cooley's data the median Series B pre-money valuation went from $152.5M in Q1 2026 to $185.2M in Q2, a 21% swing in a single quarter, and the share of deals priced above $100M pre-money rose from 41% to 48%. Up rounds ran at 83.6% of deals in Q2, flat at 4.3% and down at 12.1%. A quarter-to-quarter move that size means the comparable you anchor on in January may not be the comparable that applies in April. Anchor your ask on your own metrics and the round size you can deploy, and let the price come from the competition you have created.

If the round is not the answer, liquidity might be. Not every capital need is a priced round. Carta administered 71 tender offers in the first half of 2026 worth roughly $3 billion, the highest H1 figures in at least six years, with transaction count up 34% year over year and value up 200%. If your problem is that the team has been building for six years without a payday rather than that the company needs cash, a structured secondary is now a normal instrument rather than an admission of anything. Section 8 covers the case where a sale beats a round.

Two failure modes account for most broken processes. The first is starting before the material is ready, which turns your best-fit investors into your practice meetings. The second is letting the process drift past twelve weeks, at which point firms that passed early start hearing that you are still out, and being visibly still out is its own signal. Set the window, hold it, and if the round is not coming together by week ten, stop, go back to the metric, and return when it has moved.

11. The checklist before you open a process

Work backwards from the row that applies to you. If a Series A needs $2M of ARR and you are at $700K growing 8% a month, you are fourteen months out, which means you need eighteen months of runway and you should be raising an extension now rather than a Series A in the spring. The founders who cleared the bar in 2026 did not out-pitch anyone. They arrived with a number that made the decision obvious.