Choose how to finance the business

Research date: September 15, 2026. Every figure in this chapter was checked against a primary or near-primary source on or before this date. Venture data is revised upward for two to four quarters after a period closes, because deals get reported late — so H1 2026 numbers cited here will be restated slightly higher in later editions of the same reports. Treat anything older than 24 months as directional.

This chapter is not legal, tax, or investment advice. It describes market practice and cites public data. Securities law is one of the few areas where getting it wrong is expensive, sometimes criminal, and frequently not fixable later. Sections flagged [Lawyer required] mark points where a competent startup lawyer is not optional.

How to read this chapter

The conflict-of-interest problem in funding content

Almost everything written about startup funding is written by someone who profits from your decision.

• Venture capitalists publish blogs, podcasts, and "founder guides" that are, functionally, top-of-funnel marketing. A VC firm's product is capital; its customers are founders; its content exists to generate deal flow. This does not make the content false — much of it is the best available — but it systematically overstates how many companies should raise venture capital.
• Accelerators publish acceptance-rate-adjacent success statistics that are survivorship-selected by construction: they admit companies that were already likely to succeed, then report the outcomes.
• Bootstrapper media — indie-hacker newsletters, "calm company" podcasts, and the founders who sold a $10M-ARR business — has the mirror-image bias. It sells courses, communities, and books to people who want permission not to raise. Its base rates are also survivorship-selected; nobody writes the newsletter about the bootstrapped agency that plateaued at $400K and consumed a decade.
• Cap table and fintech vendors (Carta, AngelList, Pulley, Brex, Mercury) publish genuinely valuable proprietary datasets, and also want you as a customer. Carta's data in particular is excellent but reflects Carta's book — companies that use Carta skew toward US, Delaware C-corps, venture-track, and toward the stages where Carta sells. It undercounts the very earliest and the non-venture universe.
• Fundraising service providers — deck consultants, "fundraising-as-a-service" firms, grant writers, SEO content farms — have flooded the search results for every query in this chapter with confident-sounding numbers and no methodology. Several of the highest-ranking 2026 results for queries like "angel check size 2026" are AI-generated pages citing each other. This chapter avoids them where possible and flags them where used.

The honest position: most companies should not raise venture capital, and most companies that do raise it will not produce a venture-scale outcome. Both halves of that sentence matter.

Labels used

• [Verified] — a specific figure from a named dataset with a stated method.
• [Reported] — from a credible party whose method is partly undisclosed.
• [Estimate] — directionally useful, not precise.
• [Analysis] — our synthesis, not a measurement.

Medians, not means

Venture data is violently skewed. In H1 2026 the average US venture fund raised $188.1M while the median raised $10.8M (PitchBook-NVCA Venture Monitor, Q2 2026). A 17x gap between mean and median is the signature of a power-law distribution. Any "average" in venture is describing the top of the distribution, not the middle. This chapter uses medians wherever they exist and says so when it cannot.

The 2026 market in one page: the AI distortion

You cannot calibrate your own fundraise against "the market" in 2026, because there are two markets.

[Verified] Global venture funding hit $510 billion in H1 2026, more than the $440B invested in all of 2025, per Crunchbase (July 2026). But OpenAI and Anthropic alone accounted for 43% of that H1 total — roughly $217 billion between two companies.

[Verified] In the US, PitchBook-NVCA put H1 2026 deal value at $412.7B across roughly 9,646 deals, with AI companies capturing 86% of all venture dollars. Rounds of $100M+ made up 87.5% of deployed capital.

[Verified] Meanwhile Carta counted 4,859 total priced rounds on its platform in 2025 — a six-year low, down about 41% from the 2021 peak. Q1 2026 on Carta was $30.4B across its book.

Put those together and the shape is clear: record dollars, fewer deals. Capital is not broadly available; it is narrowly available at enormous size to a small set of companies, most of them AI-labelled. Cooley's Q1 2026 venture financing report recorded only 165 disclosed financings — "the lowest level since Q3 2016" — totalling $39.9B.

[Analysis] For a founder outside the AI capital vortex, the practical consequence is that headline valuations and round sizes in the press are useless benchmarks. The median non-AI company is raising into a market that is tighter than 2021 on deal count, with a higher metrics bar, even though the aggregate dollar figures look euphoric. The AI premium is measurable: Carta found AI startups raised Series A rounds at valuations 38% higher than non-AI peers in 2025, widening to a 193% premium at Series E+. PitchBook put median Series D+ pre-money at $4.25B for AI companies vs $134M for non-AI in H1 2026 — a gap so large it should be read as two separate asset classes sharing a label.

One more structural fact that governs everything in Part B: the exit drought is not over for the median company, even though aggregate exit value is at a record. H1 2026 US exit value was $2.19 trillion across 874 deals — but that number is dominated by SpaceX's IPO at a $1.77 trillion valuation. PitchBook's own commentary notes "the distribution shortfall that has pressured LPs for years persists." When LPs are not getting cash back, funds raise more slowly, and that flows downhill to you. First-time funds raised only $3.4B across 53 vehicles in H1 2026, annualizing to the lowest level since 2016. Experienced firms took 89% of all capital raised. Fewer new funds means fewer new first-check writers.

PART A — THE INSTRUMENTS AND SOURCES

A1. Bootstrapping and customer-funded growth

How it works

You fund the business from personal savings, salary from a job or consulting, and — the part that actually matters — customer revenue collected before or near the moment you incur cost. "Bootstrapping" as a synonym for "self-funded" undersells it. The serious version is a set of working-capital techniques:

• Prepayment. Annual contracts paid up front. A customer paying 12 months in advance lends you money at 0% and pays you for the privilege; the annual-vs-monthly mix is the single biggest cash lever a bootstrapper controls.
• Deposits and milestone billing. 50% on signature, 50% on delivery converts a cash-negative project into a cash-neutral one.
• Negative working capital. Collect faster than you pay suppliers (Net-60 out, instant capture in).
• Services-funded product. A consulting practice pays salaries while the product is built on the margin. Slow on purpose.
• Deliberate constraint on fixed cost. The binding variable is the burn floor, not revenue.

Realistic amounts and terms

There is no round. The "amount raised" is whatever gross margin you can retain, plus whatever personal capital you are willing to put at risk. Cost of capital is zero dilution and zero interest, which is why it is theoretically dominant — and very high opportunity cost, which is why it is not always right.

Who it suits

• Businesses where the customer will pay before the product is fully built (services, vertical software with configuration, B2B tools solving an urgent cost problem).
• Markets that are not winner-take-all, where being second or fifth is a fine outcome.
• Founders who value control and optionality over expected value.
• Categories where capital does not buy speed: professional services, niche software, content, most agencies, many local businesses.

Who it does not suit

• Anything with a large, irreducible pre-revenue R&D phase: biotech, semiconductors, fusion, launch, novel materials, frontier models.
• Winner-take-most network-effect markets where a funded competitor can buy the market before you reach it.
• Businesses with genuine negative unit economics until scale (some marketplaces, some hardware).

The traps

1. Confusing profitability with a business. A consultancy that pays you $200K/year is a good job, not necessarily a company. That is a fine outcome — but be honest that you chose it.
2. Under-pricing to compensate for lack of capital. Bootstrappers routinely price below what the market bears because they lack the confidence a funded competitor's balance sheet provides. Pricing is the cheapest capital available.
3. Personal financial entanglement. Personal guarantees on leases and lines of credit, credit-card balances, and second mortgages convert business risk into household risk. A failed funded startup costs you years; a failed personally-guaranteed bootstrap can cost you your house. [Lawyer required] before signing any personal guarantee.
4. Growth that outruns cash. Profitable, fast-growing companies fail on working capital all the time — you pay for inventory or headcount now and collect in 90 days. This is the specific problem venture debt and revenue-based financing exist to solve (A11, A12).
5. Survivorship in the genre. Bootstrapper media features the exits. The distribution of bootstrapped outcomes is also a power law, just with a much fatter left tail of "it works, modestly, forever."

[Analysis] The strongest argument for bootstrapping first is not ideological, it is about price. Every month you operate without outside capital is a month you build evidence that raises your valuation and shrinks the percentage you sell. The founder who raises a seed at $6M post because they have a prototype sells 20% for $1.2M. The founder who waits nine months, reaches $40K MRR, and raises at $20M post sells 10% for $2M. Bootstrapping is often best understood as valuation arbitrage, not as an alternative religion.

A2. Friends and family

How it works

Small amounts from people who know you, usually on a SAFE or convertible note, occasionally as a straight loan or gift. In the US, this is a securities offering even when it is your uncle. The exemption normally used is Rule 506(b) of Regulation D, which permits an unlimited raise from accredited investors and up to 35 non-accredited but "sophisticated" investors, with no general solicitation. Form D gets filed with the SEC within 15 days of first sale, plus state blue-sky notices. [Lawyer required] — this is cheap to do right (a few thousand dollars) and expensive to do wrong (rescission rights, meaning the investor can demand their money back, potentially years later, and it becomes a diligence problem in every future round).

Realistic amounts

[Estimate] $10K–$150K total is the common band; individual checks of $5K–$50K. Above roughly $250K you are running an angel round and should treat it as one.

Cost

Usually a SAFE with a cap that is, in practice, a favor — often $2M–$8M post-money. The real cost is not dilution. It is that these people cannot absorb the loss and cannot help you, and you will still owe them a Thanksgiving conversation.

The traps

1. Taking money someone needs. The only defensible friends-and-family raise is from people for whom a total loss is an inconvenience. Say the words "this will most likely go to zero" out loud, in writing, and get them to acknowledge it.
2. Handshake deals with no paper. The most common failure. Five years later, a verbal promise of "10% of the company" surfaces during Series B diligence.
3. Non-accredited investors on the cap table. Legal, but it complicates later rounds and some later investors will ask you to clean it up. Keep the count low.
4. Common stock sales. Selling common stock cheaply to friends creates a 409A valuation problem and can taint employee option pricing. Use a SAFE or note.
5. Rolling it up. Ten separate SAFEs with ten different caps is a cap-table mess. Use one template, one cap, one closing date if you possibly can.

A3. Grants: non-dilutive government capital

Grants are the only capital on this list that costs neither equity nor interest. They cost time — typically 3–9 months from decision to cash — and they impose reporting and compliance burden. For deep-tech companies with long pre-revenue R&D, they are frequently the single best source of early capital. For a B2B SaaS company that could ship in eight weeks, the grant application is usually a worse use of a quarter than building.

A3.1 SBIR/STTR — the big one, and it just changed

The Small Business Innovation Research (SBIR) and Small Business Technology Transfer (STTR) programs require eleven federal agencies with large extramural R&D budgets to set aside a percentage for small businesses. Roughly $4B/year flows through them.

The 2025–26 disruption matters and is under-reported. The programs' authorization expired September 30, 2025, and lapsed for six months, halting new awards across agencies. They were reauthorized on April 13, 2026, through September 30, 2031, via the Small Business Innovation and Economic Security Act (Crowell & Moring client alert, April 2026). If you are reading award data or timelines from 2025, there is a six-month hole in it.

What the reauthorization changed:

• Strategic Breakthrough Awards — a new, very large tier: individual awards up to $30 million at agencies spending over $100M/year on SBIR, with a 100% private or non-SBIR matching-fund requirement, a 48-month maximum performance period, and a 90-day contract execution requirement.
• Proposal caps to curb "SBIR mills." From FY2027, agencies must set their own limits on proposals per company, per solicitation, or per topic. Companies that have built a business on submitting dozens of proposals a year should model this now.
• Foreign risk screening. Mandatory due diligence on foreign affiliations, investment ties, technology licensing and business relationships with countries of concern; watch-listed applicants excluded. [Analysis] For startups with Chinese LP money, a foreign parent, or offshore R&D subsidiaries, this is now a real eligibility question, not a formality.
• Phase III transition support. SBA is directed to advocate for "maximum practicable use and transition" to Phase III (the commercialization phase, which has no dollar cap and can be sole-sourced).

Standard structure:

NSF (America's Seed Fund) publishes the cleanest numbers. Per the NSF 26-510 solicitation:

• Phase I: up to $305,000 over 6–18 months.
• Phase II: up to $1,250,000, typically 24 months.
• Fast-Track: up to $1,555,555 total ($400K Phase I over 6–12 months, then up to $1,155,000 Phase II).
• Phase IIB supplement: $50,000–$500,000, requiring matching investor or customer funds.
• TECP supplement: up to 20% of the Phase II award.
• Strategic Breakthrough Awards: up to $30,000,000 by Program Officer invitation for Phase II awardees.
• Process: a mandatory Project Pitch first — a short submission that gets you invited (or not) to submit a full proposal. NSF deadlines recur; the published cycle includes July 27, 2026; November 4, 2026; March 4, 2027; July 7, 2027, settling into the first Wednesday in November and first Thursday in March.
• PI requirement: the principal investigator must be at least 51% employed by the company — this disqualifies the common arrangement where a university professor is nominally the PI.

Agency character differs enormously, and this is the single most useful thing to know:

• NSF — agency-agnostic on application; funds deep tech broadly; explicitly commercialization-focused; grants (you keep IP; no deliverable to the government).
• NIH — the largest SBIR budget; therapeutics, devices, diagnostics; larger-than-standard awards routinely approved; three standard receipt dates a year.
• DoD (AFWERX, DIU, Army, Navy) — contracts, not grants; requires a government customer or "sponsor"; Direct-to-Phase-II is common; fastest path to revenue but you are building what they want.
• DOE — energy technologies; see below.

Traps:

1. Grant-funded R&D is not product-market fit. A company with four SBIR awards and no customers has learned to win grants. Investors read a long SBIR history without commercial revenue as a negative signal.
2. Timeline. Six to twelve months from submission to cash. You cannot use SBIR to fix a runway problem.
3. Compliance overhead. Cost accounting, time tracking, audits (especially on DoD contracts). Budget for an accountant who has done it before.
4. Data rights and IP. Grants (NSF/NIH) generally leave IP with you under Bayh-Dole; DoD contracts carry government purpose rights. [Lawyer required] before your first DoD contract.
5. Ownership/affiliation rules. SBIR eligibility requires >50% US ownership by individuals or certain VC/PE/hedge entities (with agency-specific limits), and generally ≤500 employees including affiliates. A single large VC owning a majority can kill eligibility. Model this before a round.

A3.2 DOE and ARPA-E

ARPA-E funds high-risk, high-reward energy R&D. OPEN solicitations run periodically; focused programs run continuously. [Verified] ARPA-E's SCALEUP program, which funds commercial scale-up for technologies that already cleared ARPA-E R&D, was succeeded in 2026 by SCALEUP Ready, announced as a rolling program — the first two projects were announced with up to $40 million committed (ARPA-E, 2026). The rolling structure is a meaningful process change: it removes the "wait 18 months for the next FOA" problem.

[Estimate] Typical ARPA-E OPEN awards run $500K–$10M with cost-share requirements that scale with technology maturity (often 5–20% for research, up to 50% for demonstration/deployment). DOE's Office of Clean Energy Demonstrations and loan programs operate at a different order of magnitude and are covered in A15.

Trap: DOE cost-share is real money you must produce, and "in-kind" treatment is narrowly defined. Confirm the cost-share percentage and allowable forms before you budget.

A3.3 EU: Horizon Europe and the European Innovation Council

The EIC is the startup-facing arm of Horizon Europe and it has a structure no US program matches: blended finance — grant plus direct equity from the EIC Fund.

[Verified] Per the EIC 2026 Work Programme, over €1.4 billion for 2026:

How the Accelerator actually works: you can apply for grant-only, equity-only, or blended. The equity is taken by the EIC Fund, typically for a minority stake, and is often the sticking point — deployment has historically been slow and the Fund's process runs on a different clock than a private round. The application is a long-form, multi-stage process (short application → full proposal → jury interview) with success rates that have run in the low single digits to low teens depending on cut-off. [Analysis] The grant is genuinely excellent money for European deep tech. The equity component requires patience and an understanding that you are adding a governmental shareholder with its own reporting requirements. The 2026 addition of STEP Scale Up and a defence-specific €30M equity line is a direct response to European scale-up capital scarcity and to the post-2022 defence-tech shift.

Trap: EU grant accounting is genuinely burdensome — audited cost claims, personnel time-sheets, consortium agreements where a partner's non-compliance becomes your problem. Budget for a grant administrator. Many European deep-tech companies use a specialist consultancy that takes 3–8% of the award; that is usually money well spent, but read the success-fee clause carefully.

A4. Accelerators as a source of capital

How it works

You give up equity in exchange for a fixed investment, a 10–14 week program, mentorship, and — the part that actually has value — a demo day and an alumni network that functions as a warm-intro machine.

Current terms at the two largest programs

Y Combinator (official deal page):

• $125,000 on a post-money SAFE for 7% of the company.
• $375,000 on an uncapped SAFE with an MFN provision, which takes the terms of the lowest-cap SAFE (or otherwise most favorable terms) issued between the start of the batch and the next equity round.
• YC also takes pro rata rights in subsequent rounds.
• No fees.

Techstars (investment terms, updated April 17, 2025):

• $220,000 total, restructured in 2025 to mirror YC's shape.
• $20,000 Convertible Equity Agreement converting to 5% common stock (post-money basis) upon a priced round of at least $1M.
• $200,000 uncapped MFN SAFE ($100,000 in Asia-Pacific programs).
• Side letter with pro rata, drag-along, and information rights.

The real cost, which founders consistently miscalculate

The headline is "7% for YC." The actual cost depends entirely on what happens to the uncapped MFN SAFE.

YC's own illustration: the $375K MFN SAFE, if the next round prices at a $15M valuation, converts to roughly 2.5%. So the true YC cost at a $15M round is roughly 9.5%, not 7%. If your next round prices at $30M, that $375K converts to ~1.25% and the total is ~8.25%. If you raise a seed at a $8M cap immediately after the batch, the MFN SAFE takes that $8M cap and converts to ~4.7%, for a total of ~11.7%.

[Analysis] The MFN structure creates an incentive you should notice: the worse your post-batch valuation, the more of your company the accelerator gets. This is not predatory — it is the standard way uncapped MFN instruments behave — but it inverts the usual intuition that an uncapped SAFE is founder-friendly. An uncapped MFN SAFE is founder-friendly only if you expect to raise at a high price.

Who it suits, and who it does not

Suits: first-time founders without a network, especially outside the Bay Area, New York, or London; teams that need the forcing function of a demo-day deadline; companies whose gap is credibility rather than product.

Does not suit: founders who already have the network (if you can get five warm intros to seed funds yourself, you are paying 7–12% for something free); companies with real revenue and real leverage, since the equity price is fixed regardless of traction, so the more traction you have the worse the deal; and, marginally, deep tech, where 12 weeks does not compress a four-year development curve.

The traps

1. Tier matters enormously and the tail is long. There are hundreds of accelerators. Outside roughly the top 10–20, the equity price is the same and the network value is close to zero. A program taking 6–10% for $50K and no meaningful demo day is a bad trade.
2. Equity-for-services programs. Some programs charge a fee and take equity. Read carefully.
3. Demo-day valuation inflation. The demo-day environment is engineered to create competitive pressure. Companies routinely raise at caps they spend two years growing into. See A6 for why a too-high cap is a real cost, not a win.
4. Corporate accelerators. Frequently structured as business development wrapped in an accelerator. They can be excellent (access to a real customer) or a time sink with an equity price attached.
5. [Analysis] Survivorship. "Companies that went through X have raised $Y billion" is selection, not causation. The relevant counterfactual — what would the same companies have done without the program — is unknowable, and the strongest programs are the most selective, which is precisely what makes the statistic uninformative.

A5. Angel investment and angel syndicates

How it works

Individuals investing personal money, typically on the same post-money SAFE everyone else uses. Angels come in roughly four species:

1. Operator angels — current or former founders and senior operators, $5K–$50K. The most useful category: they make introductions, take a support call, and judge you on what matters.
2. Professional/super angels — angel investing as a primary activity, often with a small fund. $25K–$250K.
3. Wealth angels — high-net-worth individuals with no operating background. Expect more hand-holding and more anxiety in down periods.
4. Angel groups — organized networks (Band of Angels, Tech Coast Angels, Golden Seeds, Keiretsu) pooling $100K–$1M from 10–40 members. Slow (6–12 weeks), and some charge presentation fees, which is a yellow flag.

[Estimate] The practical modern range for an individual angel check is $5,000 to $100,000, with a typical operator angel writing $10K–$25K. Figures circulating in 2026 SEO content claiming much higher "average" angel checks generally conflate angels with syndicates and with micro-funds; treat them skeptically.

Angel syndicates

A syndicate is a lead investor who finds a deal, negotiates allocation, and raises a special-purpose vehicle (SPV) from backers who invest deal-by-deal. AngelList is the dominant infrastructure; Sydecar, Allocations, and Assure-successors compete.

Mechanics that matter to you as a founder:

• One line on the cap table. The SPV is a single entity. This is the main advantage — 80 investors become one signature.
• Lead carry. The syndicate lead typically takes 15–20% carried interest on the SPV's gains, plus a small admin fee to the platform. You do not pay this; the syndicate members do.
• Speed. A syndicate lead who has done it before can close in days.
• Information rights. The SPV may pass your updates to dozens of LPs you have never met. Confirm what the SPV discloses and to whom. [Lawyer required] if you have confidentiality-sensitive metrics.

The traps

1. Party rounds with no lead. Twenty angels at $25K each and nobody who owns the relationship means nobody to call when things go wrong, nobody to anchor your Series A, and nobody who will take a board seat or do a reference call for you.
2. Signaling. A syndicate full of unknown names is neutral. A well-known micro-VC passing on your round after taking three meetings is a negative signal that propagates fast.
3. Angels who want control. An angel asking for a board seat, protective provisions, or a right of first refusal on a $25K check is a problem. Decline politely.
4. Advisory-shares-instead-of-cash. A "well-connected" angel offering introductions for 1% is almost always a bad trade. Cash checks or standard advisor agreements with vesting (0.1–1.0%, per Chapter 5), not both.

A6. SAFEs — the default early instrument, and the dilution founders get wrong

The Simple Agreement for Future Equity is a contract: the investor gives you money now, and receives equity later when a priced round happens (or cash/stock on a liquidity event). It is not debt. There is no interest, no maturity date, and no repayment obligation. [Verified] On Carta, $10.4 billion was raised across 50,316 SAFEs and convertible notes in 2025, and "the post-money SAFE with a valuation cap but no discount continues to be the standard pre-seed instrument" (Carta, State of Pre-Seed 2025).

Pre-money vs post-money: the distinction that costs founders points

YC published the original SAFE in 2013 (pre-money) and replaced it with the post-money SAFE in 2018 (YC documents). The names describe what the valuation cap measures.

• Pre-money SAFE: the cap is applied to the company's value before the new money and before the other SAFEs convert. All SAFE holders convert together and dilute each other. The founder's dilution from any single SAFE is uncertain until the round prices.
• Post-money SAFE: the cap is applied to the value after all SAFE money is counted (but before the new priced-round money). The SAFE holder's ownership percentage is fixed and knowable the day they sign: investment ÷ post-money cap.

YC's stated rationale is transparency — you can "calculate immediately and precisely how much ownership of the company has been sold." That is true, and it is also the reason the post-money SAFE shifted economics toward investors: under a post-money SAFE, every subsequent SAFE you sell dilutes you and the priced-round investors, but not the earlier SAFE holders. Under the old pre-money SAFE, earlier SAFE holders shared that dilution.

Dilution stacking — worked example

This is the single most commonly underestimated mechanic in early-stage finance, so here it is explicitly.

A founder raises across 14 months, never doing a priced round, using post-money SAFEs:

The founder raised $2M and has already sold 20% of the company before a single priced round. Many founders, thinking in terms of "I raised $2M on a $15M cap," mentally book this as ~13%. It is 20%, and it is contractually locked.

Now the seed round: a fund invests $3M at a $20M post-money valuation → 15%. The investor also requires a 15% post-close option pool.

Post-money SAFEs convert at their caps and take their fixed 20%. The new investor takes 15%. The option pool takes 15%. What is left for founders and any pre-existing common holders is:

100% − 20% (SAFEs) − 15% (new money) − 15% (pool) = 50%.

Two founders who split equally now hold 25% each, having raised $5M total. That is an entirely normal outcome and it surprises people every single time.

Three compounding effects to internalize:

1. Post-money SAFEs do not dilute each other. Sell four SAFEs and the percentages simply add. There is no "they'll all dilute each other in the wash."
2. The option pool usually comes out of pre-round ownership (the "option pool shuffle," Part B6). It is in practice a second, hidden dilution.
3. The SAFEs convert at their caps, not at the round price. If you raised on a $5M cap and price your seed at $20M, that early investor's $250K buys 5% of a company now worth $20M — they got a 4x paper markup for waiting, and you paid for it.

Valuation caps and discounts

• Cap only — the standard. The SAFE converts at the lower of the cap or the round price.
• Discount only — converts at a percentage discount (commonly 10–25%, with 20% most common) to the priced round. No cap means no ceiling on the investor's price, which is founder-friendly if your valuation rises sharply.
• Cap and discount — the investor gets the better of the two. More common in convertible notes than modern SAFEs. YC's standard post-money documents deliberately do not include a cap-and-discount variant; the three US variants are cap-only, discount-only, and uncapped MFN.
• MFN (Most Favored Nation) — the uncapped SAFE automatically adopts the most favorable terms given to any subsequent SAFE issued before the priced round. See A4 for how this behaves.

[Verified] Carta's 2025 pre-seed data: median post-money caps were around $10M for rounds of $250K–$1M and $15M for rounds of $1M–$2.5M. Caps rose across all deal sizes in 2025. AI companies sit well above these medians; non-AI companies sit at or below.

Pro rata side letters

YC's documents include an optional Pro Rata Side Letter giving the SAFE holder the right to maintain their percentage in the next round. [Analysis] Granting pro rata to every small angel is a mistake many founders make reflexively. It gives away allocation in your hot Series A — allocation your lead investor will want — and creates awkward conversations. Grant pro rata to investors who materially helped; decline it for $10K checks, politely and as policy.

Who SAFEs suit

• Pre-seed and seed raising below roughly $3–4M, where the legal cost and negotiation time of a priced round is not justified. Carta found "the majority of early-stage rounds under $4 million" used convertibles or SAFEs in 2025.
• Rolling closes: you can sign SAFEs one at a time as investors commit, rather than coordinating a simultaneous closing.
• Companies with no reasonable basis for a valuation.

The traps

1. Losing count. The single most common cap-table disaster. Maintain a live model of fully-diluted ownership assuming all SAFEs convert at their caps, updated the day each SAFE is signed. Carta, Pulley, and a competent spreadsheet all do this.
2. The cap you cannot grow into. A $25M cap raised in a frothy moment becomes a trap: your next round must price above it or your SAFE investors are unhappy, your employees' options are underwater relative to expectations, and you face a structurally awkward down round. A cap is a promise about your future, not a trophy.
3. Uncapped SAFEs without MFN. Genuinely founder-friendly, genuinely rare, and only accepted by investors who trust you completely or are buying optionality on a very hot company.
4. Side letters you forget. Information rights, major-investor status, and board observer rights granted in a side letter follow you into every future round. Keep a single side-letter register.
5. Non-standard SAFEs. An investor who hands you a "SAFE" that has been modified — a participation right, a redemption feature, an interest accrual, a most-favored-nation clause that also covers future priced rounds — is handing you a note or worse under a friendly name. [Lawyer required] for any SAFE that is not the unmodified YC form.
6. The QSBS clock. In the US, Qualified Small Business Stock holding periods generally start when stock is issued, not when a SAFE is signed. This is a real, large tax consequence for founders and early investors. [Lawyer/CPA required] — and note the QSBS rules were modified by 2025 federal tax legislation; confirm current thresholds and holding-period tiers with a tax professional before relying on any summary, including this one.

A7. Convertible notes

A convertible note is a loan that converts to equity on a qualified financing. It predates the SAFE and still has specific uses.

Mechanics

• Principal — the investment amount.
• Interest — typically 2–8% annually, usually accruing rather than paid in cash, and converting into equity along with principal.
• Maturity date — commonly 12–36 months. At maturity, absent conversion, the note is legally due. In practice it gets extended, but the theoretical right to demand repayment (or, in aggressive notes, to convert to common at a fixed price, or to force a liquidation) is real leverage.
• Valuation cap and/or discount — same as a SAFE.
• Qualified financing threshold — the minimum round size that triggers automatic conversion, commonly $1M–$5M.
• Change of control provisions — what happens if you are acquired before conversion. Commonly 1x or 2x the principal, or conversion at the cap, at the holder's election. Read this clause. A 2x change-of-control multiple on $2M of notes is $4M off the top of a $15M acquisition.

[Verified] Carta notes that convertible notes "showed greater cap fluctuation and typically carry lower valuations than SAFEs" at pre-seed (Carta, 2025).

Who they suit

• Non-US companies. SAFEs are a US-law instrument. In several jurisdictions, convertible loan notes are the established local form with clearer tax and corporate-law treatment. YC publishes post-money SAFEs only for the US, Canada, Cayman, and Singapore.
• Bridge financing where existing investors want the protection of a maturity date and seniority in liquidation.
• Investors who require debt treatment for fund or regulatory reasons.

The traps

1. Maturity in a bad market. Notes maturing during a fundraising freeze put the noteholder in control of your company's fate. This happened at scale in 2023–24.
2. Stacked notes with different terms. Three notes with three caps, three discounts, three maturities, and two different qualified-financing thresholds is a genuine legal mess to convert.
3. Interest compounding into extra dilution. 8% on $1.5M over 30 months is ~$300K of additional principal converting at your cap.
4. Debt on the balance sheet. It is a liability until it converts, which affects covenants, some grant eligibility, and how a lender views you.

A8. Priced equity rounds: seed through Series D

How a priced round actually works

You and a lead investor agree a pre-money valuation. The investor wires money; the company issues a new class of preferred stock with rights defined in a set of documents — Stock Purchase Agreement, Amended and Restated Certificate of Incorporation, Investors' Rights Agreement, Voting Agreement, and Right of First Refusal/Co-Sale Agreement. The NVCA publishes free model versions of all of them, and most US venture deals are a negotiation over deviations from those forms.

post-money = pre-money + investment. investor ownership = investment ÷ post-money. Then subtract the option pool and the converting SAFEs/notes (Part B8).

[Lawyer required] — unambiguously. Expect $25,000–$60,000 in company-side legal fees for a first priced round, plus a capped contribution to investor counsel (commonly $25K–$50K, negotiable, paid out of the proceeds).

Current market data (2025–26)

These are the numbers to calibrate against. Note the divergence between PitchBook (all US venture, includes mega-deals) and Carta (its own platform, skews to the middle of the market).

[Verified] PitchBook-NVCA, H1 2026 US medians (Venture Monitor Q2 2026):

[Analysis] Do not anchor on these. A $64M median pre-money at seed is a number produced by an AI-dominated market and by PitchBook's stage definitions. The median non-AI seed company is not raising at $64M pre. PitchBook's own split shows median Series D+ pre-money at $4.25B for AI vs $134M for non-AI — the non-AI figure is 3% of the AI figure. Apply the same skepticism at every stage.

[Verified] Cooley, Q1 2026 (venture financing report) — median pre-money for Series B was $165M (down from $195M in Q4 2025), Series D+ was $2.5B (up from $1B), and 42% of all deals across stages had pre-money valuations above $100M. Median deal size was $19.3M in technology and $22.2M in life sciences.

[Verified] Carta, 2025 full year (State of Private Markets): $119.5B raised across 4,859 rounds (six-year low in count). Average Q4 round size $30.2M, up from $19.3M a year earlier — the clearest single illustration of "fewer, larger."

Dilution per round

[Verified] Carta reported that median dilution on all rounds from seed through Series C fell from about 18% to 16% during 2025, with Series B showing the steepest decline, from roughly 15% to 12.9%.

[Verified] Carta's Founder Ownership Report gives the cumulative picture — median collective founding-team ownership:

[Analysis] Read that table carefully. It includes the effect of option pools, SAFE conversion, and multiple founders. By Series B the median founding team — often two or three people — holds under a quarter of the company, collectively. A solo founder might hold 23%; each of three co-founders might hold under 8%.

What each stage is buying

Time between rounds

[Verified] The median wait between funding rounds across all stages reached 696 days in Q2 2025 on Carta — roughly 23 months, up 5% year over year (Carta Q2 2025). Plan runway against ~24 months between rounds, not the 12–18 months of 2021 folklore.

Graduation rates — the number that should govern your planning

[Verified] Historically, 25–30% of seed-stage companies raised a Series A within 24 months. For the 2022 seed cohort, only 17% did (Carta).

[Analysis] This is the most important base rate in this chapter. Raising a seed round does not mean you are on a venture track; it means roughly one chance in four to six that you reach the next station on that track within two years. Everything in Part B about runway, milestones, and readiness follows from this one statistic.

A9. Venture capital: fund economics, and why they determine what a VC needs from you

Founders negotiate with VCs as if the VC's goal is "a good return on this investment." It is not. The VC's goal is to return their fund, and the structure of a venture fund makes that goal achievable only through extreme outliers. Understanding the arithmetic explains almost every behaviour founders find confusing.

The structure

A venture fund is a limited partnership. Limited partners (LPs) — pensions, endowments, foundations, sovereign wealth funds, funds-of-funds, family offices — commit capital. General partners (GPs) — the VCs — invest it.

• Term: typically 10 years, with two or three one-year extensions. The fund must eventually liquidate. This is why your investor cares about a timeline even if you do not.
• Investment period: roughly the first 3–5 years; reserves are held for follow-ons after that.
• Management fee: ~2% of committed capital per year (often stepping down after the investment period). A $100M fund spends roughly $15–20M on fees over its life, meaning only ~$80–85M is actually invested.
• Carried interest: typically 20% of profits above return of capital, sometimes above a preference hurdle. Elite firms take 25–30%.

The arithmetic of a fund return

Take a $100M fund. To be a "good" fund it should return 3x gross — $300M — because after fees and carry that produces roughly a 2.5x net to LPs, which beats public-market alternatives enough to justify the illiquidity.

It invests ~$85M across, say, 30 companies, averaging ~$2.8M each with reserves. For $300M back:

• If the best company returns 30% of the fund — $30M — it is not enough. You need ten of those.
• The empirical reality is that one or two companies return most of the fund. A 20% ownership in a company that exits at $1B returns $200M — two-thirds of the target from a single position.

This is the power law. Carta's VC Fund Performance report, covering 2,775 funds from 2017 through Q1 2026 (~$119.3B), makes the consequence visible:

[Verified]

• Top-decile net IRR exceeds 20% for every vintage 2017–2024 except 2021 — but the 75th percentile is below 15.5%. The distribution is not a bell curve with a long tail; it is a spike.
• For 2019 and 2020 vintages, median DPI barely exceeds zero — fewer than half of those funds have returned any capital to LPs.
• Fewer than 20% of 2017 and 2018 vintage funds have reached 1x DPI — that is, have returned even the money LPs put in, eight and nine years after inception.
• Median TVPI for 2018 funds was 1.38x; 90th percentile was 3.07x (Carta Q3 2025).
• Median IRR for the 2021 vintage was 0.5%; for 2022, 0.1%.

[Analysis] The median venture fund is a mediocre investment. That is not a scandal; it is the known shape of the asset class, and LPs accept it because access to the top decile is worth the drag. But it explains everything:

1. Why your VC needs a huge outcome from you specifically. A $50M acquisition that triples your investor's money is a great outcome for you and an irrelevant one for a $500M fund. This is the source of the most common founder-investor conflict.
2. Why VCs push growth over profitability. Profitable-and-modest is a fine business and a poor fund position.
3. Why VCs pass on good companies. "This is a real business but I can't see the $1B outcome" is not an insult; it is a portfolio-construction statement.
4. Why fund size matters more than brand. Ask any prospective investor: how big is this fund, what's your target ownership, what does a good outcome for this position look like? A $30M fund is thrilled with a $100M exit. A $2B fund is not. This one question is the most useful piece of diligence a founder can do on an investor, and almost nobody asks it.
5. Why the current LP distribution drought is your problem. Carta's Q1 2026 framing is blunt: managers "will need to convert the unrealized gains into concrete returns" to sustain the cycle. When LPs are not getting cash, they commit less, funds shrink, and fewer new funds form — first-time funds raised only $3.4B across 53 vehicles in H1 2026, annualizing to the lowest since 2016 (PitchBook-NVCA).

Fund types you will encounter

[Analysis] The multi-stage seed check has a hidden cost. When a $5B fund writes you a $1M seed check, that position is immaterial to them but highly material to you. If they decline to lead your Series A, the market reads it as an informed pass — the strongest negative signal in venture. A dedicated seed fund that must mark up its position has interests better aligned with yours at that moment.

A10. Corporate venture capital

How it works

A corporation invests in startups, either from the balance sheet or via a dedicated fund vehicle with an independent-ish team. The objectives are typically strategic (market intelligence, ecosystem building, a pre-M&A look) as much as financial.

Current scale

[Verified] Corporate investors accounted for a record 87.9% of US AI venture deal value in 2026 year-to-date, per PitchBook's Q3 2026 corporate venture report, while their share of deal count has consistently declined. AI now makes up over 90% of all corporate VC deal value.

[Analysis] That 87.9% figure is not "corporates are everywhere." It reflects a handful of enormous strategic transactions — hyperscalers investing in frontier labs — dominating a value-weighted statistic. The number of corporates writing normal-sized checks has contracted from the 2021 peak. For a typical Series A company, CVC is less available than the headline suggests.

PitchBook's taxonomy of corporate motives is useful:

• Cloud hyperscalers (Amazon, Microsoft, Google) invest to "secure model access and lock in cloud commitments" — the investment and the commercial contract are the same decision.
• Enterprise/infrastructure firms (Salesforce, Cisco, Qualcomm, Intel) invest to embed capability in their own products.
• NVIDIA backs competing labs to preserve demand for its chips.
• Frontier labs deploy capital as a go-to-market strategy.

Amounts and terms

[Estimate] $250K–$25M depending on the corporate and stage. Most CVCs take the same preferred security as the financial lead and follow standard terms. The negotiation is over the strategic terms.

Who it suits

• Companies selling into the corporate's industry, where the investment unlocks distribution or a lighthouse customer.
• Deep tech and hardware needing manufacturing, supply chain, or regulatory access money cannot buy.
• Capital-intensive sectors (energy, health systems, automotive) where a corporate's balance sheet is patient in a way a 10-year fund cannot be.

The traps

1. Right of first refusal / right of first offer on acquisition. The single most dangerous CVC term. A ROFR on an acquisition chills every other bidder — why run a process if one party can match at the end? Refuse it. If you cannot, cap it tightly in time and scope. [Lawyer required]
2. Exclusivity and most-favored-customer clauses. An investment agreement that also makes you exclusive to one distributor can shrink your market by 80%.
3. Signaling to the corporate's competitors. Taking Salesforce money can make selling to Microsoft harder. In some sectors this is fatal; in others nobody notices. Know which you are in.
4. Information rights to a potential competitor. Board observer seats give a strategic investor a quarterly view into your metrics, roadmap, and pipeline. Scope observer rights explicitly and reserve the right to exclude them from competitively sensitive sessions.
5. Strategic champion risk. The executive who sponsored your investment leaves, the strategy changes, and your "strategic partner" becomes a passive shareholder with awkward rights. Assume this will happen. Never build a plan that depends on the corporate's continued enthusiasm.
6. Slow. CVC processes run on corporate legal and procurement timelines. Add 4–10 weeks.

A11. Venture debt

How it works

A loan to a venture-backed company, underwritten primarily on the quality of your equity investors and your ability to raise again, not on assets or cash flow. Usually taken within 3–12 months of an equity round, when your balance sheet looks strongest.

The post-SVB landscape

Silicon Valley Bank's March 2023 failure removed the institution that had underwritten a large share of US venture debt. The market did not disappear; it redistributed and repriced.

[Reported] The current lender set (startupfundraising.com venture debt playbook, a fundraising-services site — treat terms as indicative, not audited):

• Banks: SVB (now part of First Citizens), HSBC Innovation Banking (which acquired SVB UK), JPMorgan, Bank of America, Comerica, PNC, Bridge Bank / Western Alliance. Stifel Venture Banking also built a franchise from ex-SVB teams. Rates around SOFR + 6–10% (roughly 10–14% all-in), warrant coverage 5–15%, tighter covenants.
• Non-bank/BDC lenders: Hercules Capital, TriplePoint Capital, Runway Growth Finance, Trinity Capital, Horizon Technology Finance. Rates 11–15%, warrant coverage 10–25%, more flexible when things go wrong.
• Life sciences specialists: Oxford Finance, and the life-science arms of Hercules and Runway.

Typical 2026 structure:

• Term: 36–48 months, with 12–24 months interest-only before amortization.
• Size: commonly 20–35% of the most recent equity round, or a multiple of ARR (often 0.3x–0.6x ARR) for revenue-based structures.
• Fees: 0.5–1.5% upfront commitment fee, 0.5–1% back-end/final payment fee, plus legal and expense reimbursement often $50K–$150K.
• Prepayment penalties: roughly 2–3% in year 1, 1–2% in year 2, 0% thereafter.
• Covenants: minimum cash (often 3–6 months of runway), material adverse change (MAC) clauses, and ARR or revenue covenants at later stages.

The real cost

Interest is the advertised price. Warrants are the hidden one. Warrant coverage is expressed as a percentage of the loan amount: 10% coverage on a $10M loan means the lender receives warrants to buy $1M of stock, usually at the price of your most recent round, exercisable for 7–10 years. On a company that does very well, those warrants are expensive. On a company that fails, they are worthless — which tells you the lender is pricing for the outcome where you succeed.

Who it suits

• Post-Series-A/B companies with a specific milestone in mind. Debt buys time to hit a value inflection; it does not buy survival.
• Extending runway between rounds when you are 2–4 months short of a metric that would materially raise your valuation. Borrowing $3M to avoid selling 15% more equity is often excellent math.
• Working capital — inventory, equipment, receivables — where the asset supports repayment.

It does not suit pre-revenue companies without committed equity backers, or struggling companies: debt does not fix a demand problem, it adds a fixed obligation to a company that cannot meet it.

The traps — this is the section to read twice

1. MAC clauses. A "material adverse change" clause lets the lender declare default on subjective grounds. Negotiate it out or narrow it. [Lawyer required]
2. Cash covenants that trigger exactly when you are weakest. A minimum-cash covenant means the lender can sweep your account at the moment you most need it. Model the covenant against your downside plan, not your plan.
3. Debt is senior to everything. In liquidation, the lender is paid before preferred, which is paid before you. Venture debt converts a bad outcome into a zero for common holders.
4. Amortization arrives. The interest-only period ends and your monthly cash cost jumps three-to-fivefold, often precisely when you are fundraising.
5. Investor consent. Most equity documents require board or preferred approval for indebtedness. Get it early.
6. "It's non-dilutive." It is less dilutive. Warrants, plus the increased risk of a down round or failure caused by fixed payments, are a real equity cost.

[Analysis] The one-sentence rule: venture debt is appropriate when you can articulate exactly which milestone the money buys and you would still survive if that milestone slipped by six months.

A12. Revenue-based financing and recurring-revenue advances

How it works

A financier advances cash against future revenue. Repayment is a fixed percentage of monthly revenue until a predetermined multiple of the advance is repaid (typically 1.06x–1.5x), or the provider buys your annual contracts at a discount and collects from your customers. No equity, no board seat, no personal guarantee in the standard forms, and often no fixed maturity — if revenue falls, repayment slows.

Current status of the named providers — this sector consolidated hard

[Verified] Pipe, which launched the "trading platform for recurring revenue" model and reached a $2B valuation in 2021, no longer runs that business. Per Pipe's own 2026 post, the company "relaunched and refocused in 2024 with the launch of Pipe Capital," pivoting from a standalone marketplace to embedded capital distributed through software platforms — partners named include Uber, Boulevard, Housecall Pro, Live Payments, and GoCardless. The original Pipe product a SaaS founder would have used in 2021 is gone. If you find a 2021–22 article recommending Pipe for SaaS revenue trading, it is out of date.

[Verified] Capchase remains active and has continued to raise debt facilities to fund its book (Crunchbase News), expanding from pure revenue-based advances into B2B buy-now-pay-later — financing your customers' purchases of your software rather than advancing against your revenue. This is a meaningful distinction: Capchase Pay is a sales tool (it lets your customer pay monthly while you get paid annually up front), not a loan to you.

[Reported] Founderpath continues to operate as a SaaS-focused lender, positioning itself against SVB and bank alternatives (Founderpath comparison pages — vendor-published, self-serving, cite with caution).

Others in the category: Lighter Capital (one of the oldest RBF providers, focused on $50K–$4M facilities), Arc, Levenue (Europe), re:cap (Europe), Uncapped, and Clearco (which restructured substantially after its 2021–22 peak and now focuses on e-commerce inventory and marketing spend).

[Analysis] The structural lesson of this sector. Revenue-based financing was built in a zero-interest-rate environment where the cost of the provider's own capital was near zero and the spread was easy. When base rates rose, the economics of advancing at 1.06x–1.12x against SaaS revenue compressed or inverted, and providers either raised prices, tightened underwriting, pivoted to embedded/B2B2C distribution (Pipe), or moved up the stack into products with better margins (Capchase). Any founder evaluating an RBF provider in 2026 should assume the provider itself may pivot or exit within 24 months, and should read the assignment and servicing provisions accordingly.

Realistic amounts and cost

[Estimate] Facilities of $50K–$10M, most commonly sized at 2–5x MRR or up to ~30–60% of ARR. Effective annualized cost typically 12–25%, sometimes higher once fees are annualized over a short repayment period. The advertised "flat fee of 8%" on a 9-month repayment is an APR in the high teens to twenties.

Who it suits

• Predictable, high-retention recurring revenue — SaaS above roughly $20K MRR with low churn.
• E-commerce with proven, measurable return on ad spend, where $1 in equals $2.50 out on a known lag.
• Bridging annual-contract cash flow: you sell an annual contract billed monthly and want the cash now.

The traps

1. Annualize the cost. A "6% fee" repaid over six months is roughly a 12%+ APR. Over three months, 24%+. Always compute the implied APR.
2. Revenue share, not profit share. Repayment comes off the top line regardless of margin. A 10% revenue share on a business with 25% net margin consumes 40% of your profit.
3. Stacking. Taking a second and third advance to repay the first is the fastest route to insolvency in this category, and it is common. Providers will happily let you.
4. Covenants and control. Read for lockbox arrangements (payments routed through the provider's account), assignment of contracts, and consent requirements on a sale of the business. [Lawyer required]
5. It does not buy growth capital, it buys timing. RBF is a working-capital instrument. Using it to fund an unproven expansion is how businesses with good unit economics become businesses with debt and bad unit economics.

A13. Equity crowdfunding (Regulation Crowdfunding)

How it works

Under Regulation Crowdfunding (Reg CF), a US company can raise from the general public — including non-accredited investors — through an SEC-registered funding portal or broker-dealer. The cap is $5 million in any rolling 12-month period (a figure adjusted periodically for inflation; confirm the current number with counsel before relying on it). Individual investor limits are tiered by income and net worth. Offerings require a Form C filing with financial disclosure — reviewed financials above a threshold, audited above a higher one — and annual ongoing reporting.

Related regimes: Reg A+ Tier 2 allows up to $75M/year with substantially heavier disclosure and SEC qualification (effectively a mini-IPO), and Reg D 506(c) allows unlimited raises with general solicitation but accredited investors only, with verification.

What the outcomes actually look like — the SEC's own data

This is the most useful and least-cited dataset in the category. [Verified] From the SEC Division of Economic and Risk Analysis "Analysis of Regulation Crowdfunding" (May 2025), covering May 2016 through December 2024:

• 8,492 offerings initiated by 7,134 issuers (excluding 990 withdrawn).
• 3,869 offerings (46%) reported proceeds, totalling about $1.3 billion across nearly nine years.
• Average successful raise: $346,000. Median successful raise: $113,000.
• Median issuer profile: $80,000 in total assets, 2.4 years old, 3 employees, $9,800 in revenue, −$32,400 in net income. Only ~14% had positive net income.
• Security types: equity 43%, debt 31%, SAFEs 25%.
• Platform share of initiated offerings: Wefunder 28.6%, StartEngine 19.0%, Honeycomb 9.5%, Republic/OpenDeal 8.0%; top five platforms ≈70%.
• Intermediary compensation averaged 7.7–8.1% of proceeds (cash commissions ~6.5% plus securities where taken).

And the outcomes:

• IPO: 0.25% of issuers.
• Acquired: 2.2%.
• Subsequently raised venture capital: 3.4%.

[Analysis] Read those three numbers together with the median raise of $113,000. Reg CF is not a venture-substitute for most companies that use it; it is a small-business and community-capital instrument that occasionally funds a consumer brand with a passionate audience. The marketing on platform websites emphasizes the handful of companies that went on to large rounds; the SEC's population-level data says 96% did not raise venture capital and 97.5% have not exited.

Who it suits

• Consumer brands with an existing audience — breweries, food, DTC, games, media. The raise is a marketing event and the investors become customers and evangelists. This is the genuine use case.
• Community businesses — a local restaurant, a co-op, a neighbourhood venue. Honeycomb's presence in the platform rankings reflects real debt-based community lending.
• Companies that want to reward an existing user base with upside.

Who it does not suit

• B2B software. There is no consumer audience to activate, and institutional investors are indifferent at best.
• Companies planning an institutional round soon, unless structured carefully.

The traps

1. Cap table hygiene. Thousands of direct shareholders is a diligence problem and a signature problem. Use a platform that offers a custodial or nominee structure so the crowd appears as one line. Wefunder's SPV structure and StartEngine's equivalents exist for this reason. If you take 2,000 individual holders directly onto the cap table, expect institutional investors to ask you to clean it up at your own expense.
2. The 8% take. Roughly a twelfth of what you raise goes to the intermediary before you pay for video production, ads, and legal.
3. Ongoing reporting. Annual Form C-AR filings, forever, until you deregister. Public-ish disclosure of your financials, visible to competitors.
4. The 500-holder-of-record threshold. Exceeding certain holder-of-record counts can trigger Exchange Act registration obligations. Reg CF securities have a conditional exclusion, but the conditions matter. [Lawyer required]
5. Regulatory history. FINRA fined Wefunder $1.4M and StartEngine $350,000 in 2022 for crowdfunding rule violations (BusinessWire, May 2022). The category has matured since, but portal compliance is not a given.
6. Failed raises are public. A campaign that stalls at 20% of target is visible to everyone, including future investors.

A14. Strategic investment (non-CVC)

A customer, supplier, distributor, or partner invests — often alongside a commercial agreement. Distinct from CVC in that there is usually no dedicated investing team, no fund, and no pretence of financial discipline.

Amounts: [Estimate] $100K–$10M, frequently sized to match a commercial contract.

What makes it good: the investment is a commitment device. A distributor who has written a cheque works harder on your product. A customer-investor in a long enterprise sales cycle is a reference you cannot buy.

What makes it dangerous: you can no longer tell whether your revenue is real. A strategic investor who invests $2M and then buys $1.5M of software has given you $500K and a growth chart that will not survive diligence. [Analysis] This is one of the most common reasons Series B diligence blows up: the acquirer or investor discovers that a large share of ARR comes from shareholders. Disclose it proactively, segment it in your metrics, and never let investor-customers exceed a modest share of revenue.

Also watch: exclusivity, ROFR on acquisition (same rule as A10 — refuse), MFN pricing on the commercial contract, and the "we'll invest if you build this feature for us" trade, which is a services contract wearing an investment costume.

A15. Government funding beyond grants

Grants (A3) are the visible piece. There is considerably more.

US:

• SSBCI (State Small Business Credit Initiative). [Verified] Reauthorized and expanded with $10 billion under the American Rescue Plan, allocated to states, territories, and tribal governments who design their own programs — loan guarantees, collateral support, loan participation, and direct equity/venture programs. Some states invest directly in startups, some seed local venture funds, and many programs explicitly target underserved founders. Treasury estimated the program would attract roughly $10 of private capital per SSBCI dollar (Carta explainer; NVCA SSBCI resources). Terms, availability, and stage focus vary enormously by state, and not every participating state built an equity program. This is the most underused capital source in the US for non-coastal companies.
• SBA loan programs. 7(a) and 504 loans, delivered through banks with an SBA guarantee. Generally require collateral, personal guarantees, and cash flow — which disqualifies most pre-revenue startups but fits profitable small businesses, acquisitions, and equipment purchases well. [Lawyer required] before signing a personal guarantee.
• DOE Loan Programs Office and the Office of Clean Energy Demonstrations: debt at a scale (hundreds of millions to billions) relevant only to capital-intensive energy projects.
• Defense: beyond SBIR, the Defense Innovation Unit, AFWERX/SpaceWERX, Office of Strategic Capital, and In-Q-Tel (the CIA's strategic investor) all deploy capital. DIU's Commercial Solutions Opening process can move faster than traditional procurement.
• NIST, USDA, NOAA, DOT and others run sector programs and, in agriculture and rural development, direct loan and guarantee schemes.
• State and city incentives: R&D tax credits, job-creation credits, relocation grants, and payroll tax offsets for R&D credits — the federal R&D credit can offset up to a capped amount of payroll tax for qualifying small businesses, which is real cash for a pre-profit company. [CPA required]

Europe: national innovation agencies (Innovate UK, Bpifrance, Enterprise Ireland, Vinnova, Business Finland, EXIST) offer grants typically €25K–€2M, often equity-free. In the UK, SEIS and EIS reliefs do not fund you directly but make you dramatically more investable to UK angels (SEIS gives investors 50% income tax relief); HMRC Advance Assurance before raising is standard practice. [Lawyer/accountant required] The EIB/EIF underpin much of European VC through venture debt and fund-of-funds activity.

Elsewhere: Canada's SR&ED refundable credits and IRAP; Israel's Innovation Authority; Singapore's Startup SG and EDB; Japan's NEDO; Australia's R&D Tax Incentive.

[Analysis] For a pre-revenue deep-tech company in a jurisdiction with a serious R&D credit regime (Canada, UK, France, Australia), the refundable credit can be worth 20–45% of qualifying R&D spend in cash. It is often the single largest non-dilutive line in the first three years and is routinely under-claimed because founders do not know their engineering salaries qualify.

A16. Private equity and growth equity

Growth equity

Minority investments of $20M–$200M+ into companies that are already scaled — typically $10M–$100M+ ARR, growing 30%+, at or near profitability. Investors include Insight Partners, General Atlantic, Summit Partners, TCV, Bessemer's growth practice, and the growth arms of large VCs. Structure resembles a late venture round: preferred stock, 1x non-participating preference, board seat, protective provisions.

What distinguishes it from late-stage VC: growth equity underwrites existing economics rather than future ones. The diligence is quantitative — cohort retention, CAC payback, net revenue retention, gross margin by segment — and it will find everything. A growth round is a good option for a company that has become a real business and does not need a venture outcome to make the numbers work.

Frequently includes founder secondary. This is one of the main reasons founders take growth rounds (see B15).

Private equity buyouts

Majority control, typically 60–100% of the equity, usually with leverage (debt on the company's balance sheet). Buyers include Thoma Bravo, Vista, Francisco Partners, and hundreds of mid-market and lower-mid-market funds. In software this is now a fully institutionalized exit path.

What PE pays: [Verified] Aventis Advisors' dataset of 543 private SaaS M&A transactions shows a long-run median EV/Revenue of 4.5x and median EV/EBITDA of 23.0x, with median deal size around $80M. But the recent years matter more than the long run: 2021 peaked at 6.3x revenue, 2023 fell to 3.7x, 2024 to 2.9x, 2025 recovered to 3.8x, and Q1 2026 was 3.1x (Aventis Advisors, 2026).

Size drives multiple sharply. In the same dataset: ~3.2x for $0–5M ARR, ~3.0x for $5–20M, ~3.3x for $20–50M, ~6.1x for $50–100M, ~5.1x for $100–500M, ~6.2x for $500M+. The discontinuity above $50M ARR is where institutional buyers enter and financing becomes available. [Analysis] This is the clearest argument in the data for not selling at $15M ARR if you can credibly reach $60M: the multiple roughly doubles, on top of the revenue growth.

Rule of 40 matters measurably: Aventis found a 10-point increase in Rule of 40 corresponds to roughly 1.02x additional EV/Revenue.

For comparison, public SaaS traded at a median 4.6x EV/Revenue in August 2026 — down from 17x at the 2021 peak, and a reminder that the private multiple ceiling is set by the public comparables.

The lower end: Acquire.com's January 2026 multiples report covers sub-$10M enterprise value SaaS deals and reports a median confirmed sale at 3.9x profit (not revenue), with an average 81 days on market and most deals closing within 90. Bootstrapped SaaS businesses at this scale are valued on profit, not ARR — a distinction that catches founders who have absorbed venture-market vocabulary.

The traps

1. Leverage changes the company. Post-buyout, debt service is the first call on cash. Product investment, hiring, and R&D get reprioritized accordingly.
2. Management rollover. PE buyers usually require founders/management to roll 10–30% of proceeds into the new entity. That is a second bet on the same asset, now leveraged, with someone else in control.
3. Earnouts. A headline price that includes a large earnout is not the price. Earnouts are frequently not paid in full, and the acquirer controls the conditions.
4. The process is exhausting and leaks. Quality-of-earnings reviews, management presentations, and 6–9 months of diligence, during which the business must keep performing.

A17. Acquisition as a financing and exit event

Acquisition is where most successful startups end. It is also, viewed correctly, a financing decision — you are choosing to convert future upside into present certainty.

The 2026 M&A environment

[Verified] Global M&A for venture-backed companies hit records in H1 2026: 24 acquisitions at $1B+ in Q2 alone, totalling $113 billion (Crunchbase). PitchBook put US YTD acquisition value at $375.4 billion, a decade high (Venture Monitor Q2 2026). The largest single transaction reported was SpaceX's acquisition of Anysphere (Cursor) at approximately $60 billion.

[Analysis] Again: concentration. A record aggregate driven by a handful of AI transactions does not mean the median startup has a better chance of being acquired. The mid-market — $20M to $200M software acquisitions — is a different market governed by the multiples in A16 and by strategic buyers' own budgets.

The four kinds of acquisition

1. Strategic acquisition at a premium. The buyer wants your product, customers, or market position and pays a multiple of revenue. This is the outcome everyone imagines.
2. Financial acquisition (PE). Priced on profit and cash flow. See A16.
3. Acqui-hire. The buyer wants the team. Consideration is typically $1M–$3M per engineer as a rough heuristic, but the critical structural fact is that most of the consideration is allocated to retention packages for the employees who join, not to the purchase price paid to shareholders. A $15M "acquisition" of a 12-person company may be $4M to the cap table and $11M in four-year retention grants. If you raised $10M with a 1x liquidation preference, common holders get nothing. Founders and employees may do fine personally while shareholders get wiped out, which is a structural conflict worth naming out loud with your board.
4. Distressed sale / asset sale. The company is out of runway. Buyers pay for IP and customers. Proceeds usually do not clear the preference stack.

Mechanics that determine what you actually receive

• The liquidation preference stack. Preferred is paid first. With $40M raised at 1x non-participating and a $45M sale, preferred takes $40M and common splits $5M. Model this before accepting a high valuation on a large raise (B7).
• Escrow/holdback. Typically 10–15% of consideration for 12–24 months against reps-and-warranties breaches. Representation and warranty insurance can reduce or eliminate it — worth negotiating.
• Stock vs cash. Acquirer stock in a private company is not money. Discount it and read the transfer restrictions.
• Vesting acceleration. Single- vs double-trigger acceleration determines whether your team's equity vests on the deal. Buyers dislike full single-trigger because it removes retention. [Lawyer required]
• Drag-along rights. Investors can usually force common holders to vote for a sale preferred approves. You may have less control over the decision than you think.

The traps

1. "We'd love to acquire you" is often diligence in disguise. An unsolicited approach with no price and a long information request may be a competitor learning your business for free. Require a price range and an NDA before opening the books.
2. The process destroys focus. Six months of diligence while missing your numbers is the worst possible position, because the buyer can reprice.
3. Acquisition is not a plan. "We'll get acquired" is a hope; buyers are few, budgets are cyclical, and the decision is not yours.

A18. The instruments compared

PART B — THE PROCESS AND THE DECISION

B1. When to raise, and when not to

Raise when all of these are true

1. Capital is the binding constraint. You know exactly what the money buys and what it unblocks. "We'd grow faster" is not a constraint; "we have 400 qualified leads a month and two salespeople" is.
2. You have evidence, not just a plan. Something is working and more money makes more of it happen.
3. The market rewards speed. Network effects, a land-grab, a closing regulatory window, or a well-funded competitor moving.
4. You can live with the obligations. A venture round is a commitment to pursue a very large outcome on a roughly 7–10 year clock.

Do not raise when

• You are raising to avoid the harder problem. Money does not create demand. A company without customers that raises $3M becomes a company without customers and 18 months of payroll.
• Your metrics are about to improve sharply. Waiting two quarters to raise at double the valuation halves your dilution for the same dollars.
• You do not know what the money is for. If you cannot name three specific milestones and their cost, you are not ready.
• The business does not have a venture-scale ceiling. Taking venture money into a $50M-TAM business is a trap for everyone, and you will be the one living in it.
• You would take any terms. Desperation is visible and it gets priced.

[Analysis] The base rate to keep in front of you: roughly 17–30% of seed-funded companies reach Series A within 24 months (Carta). Raising a seed round is a qualifying event, not an arrival. Plan for the 70–83% case: what does this company look like if the Series A does not happen?

B2. Fundraising readiness: what actually unlocks each stage in 2026

Every metric bar below is a rough central tendency for B2B software in the current market, not a rule, and the AI premium distorts all of them upward for companies that can credibly claim the label. Consumer, marketplace, fintech, biotech and hardware have different currencies entirely.

Non-metric readiness — the parts that kill deals in diligence:

• Clean cap table. All SAFEs modelled, all side letters catalogued, all equity grants board-approved and documented, all 83(b) elections filed (Chapter 5).
• Corporate housekeeping. Delaware C-corp in good standing (or local equivalent), board consents for every issuance, minutes that exist.
• IP assignment from everyone. Every founder, employee, contractor, and that friend who designed the logo in 2023. Missing contractor IP assignments are the single most common diligence blocker.
• Financials that reconcile. Accrual-basis books, a revenue recognition policy, and an ARR definition you can defend line by line. If your ARR number in the deck does not tie to your accounting, you will be asked about it in front of an investment committee.
• Founder alignment in writing. Vesting, roles, and what happens if someone leaves.

B3. Investor targeting and warm intros

Targeting

Build a list of 60–120 firms, tiered. For each, record: fund size, current fund vintage (a fund in year 5 of a 5-year investment period has no money), stage focus, check size, target ownership, sector thesis, portfolio conflicts, and the specific partner who would sponsor you. Partner-level targeting matters more than firm-level: firms do not invest, partners do, and each partner has a limited number of new board seats per year.

Tier A (15–25): genuine fits who would be excellent partners. Tier B (30–50): good fits. Tier C: everyone else. Run Tier B first to rehearse, then Tier A within a compressed window to create real simultaneity.

Warm intros

[Analysis] The preference for warm introductions is not snobbery; it is a filtering mechanism for people who receive thousands of inbound decks. The signal is not "someone vouched," it is "you were able to find a path," which correlates with the resourcefulness the job requires.

Ranked by strength:

1. A founder the investor has already backed and made money with.
2. Any portfolio founder.
3. A co-investor who is already committed to your round.
4. An LP in the fund.
5. A lawyer, banker, or accountant who works with them regularly.
6. Cold outbound that is genuinely excellent.

Mechanics: send a forwardable email — three to five sentences the introducer can pass on without editing, plus your deck. Always give the introducer an easy out ("no pressure at all if you don't think it's a fit"). Never CC the investor on the ask; use double opt-in.

Cold outbound works more often than folklore suggests, particularly with seed funds and solo GPs who compete on access. It works when the email demonstrates you know the specific partner's thesis, states traction in the first two lines, and asks for a 20-minute call rather than "coffee."

Process discipline: run the raise as a compressed, parallel process, not sequentially. Sequential raising creates a public record of passes. Parallel creates the only leverage a founder has — a deadline that is real because two parties are interested at once.

B4. Pitch decks

Current structure norms (10–15 slides for the deck you send; ~20 including appendix)

1. Title / one-line positioning — in the language a customer would use.
2. Problem — specific, evidenced, quantified for a named buyer.
3. Solution / product — show the thing; screenshots beat adjectives.
4. Why now — the technological, regulatory or behavioural change that makes this possible now and not in 2019. Investors weight this heavily and founders routinely omit it.
5. Market — bottom-up TAM (buyers × realistic ACV). A credible $2B beats an incredible $80B.
6. Traction — the most important slide from seed onward. Axes labelled, units stated.
7. Business model — pricing, unit economics, gross margin.
8. Go-to-market — the channel that works, and the evidence it works.
9. Competition — honest positioning. Never a 2x2 with you alone in the top-right.
10. Team — why these specific people.
11. Financials / plan — 24 months, assumptions visible.
12. The ask — amount, what it buys, which milestones it reaches.

Appendix: cohort retention, detailed unit economics, pipeline, product roadmap, customer references, org plan.

What the deck analyses actually show

[Verified, dated] DocSend's seed studies remain the most-cited empirical work, and the numbers are old enough to flag: in its 2019 study of 175 seed-stage startups, DocSend found successful founders contacted an average of 77 investors and held roughly 40 meetings to close a round; unsuccessful founders contacted 70 investors but got only ~15 meetings. Deck length for a successful seed raise was 20 slides, unchanged from 2015. Investors spent an average of 3 minutes 27 seconds on a deck (down from 3:44 in 2015).

[Analysis] Three things follow. First, the constraint is meetings, not decks — the gap between successful and unsuccessful founders was conversion to meetings, not volume of outreach. Second, under four minutes of attention means the first three slides carry the deck. Third, these figures predate the 2022 reset and the AI era; assume the required outreach volume is higher now for non-AI companies, and note that DocSend is owned by Dropbox and publishes this research to market a document-sharing product.

Beware the genre of "decks that raised $X" analyses. They are selected on the outcome, and the deck is rarely why the round happened. Uber's seed deck is famous because Uber succeeded.

B5. Data rooms

Assemble before you start, not when asked. A same-day response to a diligence request is itself a signal about how you run the company.

Contents: Corporate — charter and amendments, bylaws, board and stockholder consents, minutes, Form Ds, good-standing certificates. Cap table — fully diluted, every SAFE and note modelled at its cap, all grants, all side letters, current 409A. Financials — 24–36 months of P&L, balance sheet and cash flow, plus the model with assumptions visible. Metrics — ARR bridge (new/expansion/contraction/churn), cohort retention, CAC by channel, payback, pipeline. Customers — top-20 contracts, standard MSA/DPA, and the churned-customer list with reasons (include it; they will find it). Team — org chart, offer letters, every IP assignment and contractor agreement, the option plan. IP — patents, trademarks, open-source inventory, third-party and university licenses. Legal — litigation, insurance, leases, regulatory correspondence.

Practices: use a real data room with per-document access logs; stage access (Tier 1 for first diligence, Tier 2 after a term sheet); watermark. Who looked at what, for how long, tells you which investors are serious.

B6. Term sheets: the terms that actually matter

A term sheet is mostly non-binding except for confidentiality, exclusivity/no-shop, and expenses. It nonetheless sets everything, because renegotiating after signing is possible only at reputational cost. [Lawyer required] — a startup lawyer who does this weekly, not a generalist.

Economic terms

Liquidation preference. The multiple of invested capital preferred holders receive before common. [Verified] Market standard is overwhelmingly 1x: Cooley's Q1 2026 data shows 98.2% of deals at 1x and 96.4% non-participating (Cooley Q1 2026); Q4 2025 was 98% and 96% (Cooley Q4 2025). Anything above 1x is a structured round and should be priced as such (B13).

Participation. Non-participating: the investor chooses either the preference or their pro-rata share of the proceeds, whichever is greater. Participating ("double dip"): they take the preference and share pro rata. On a $50M exit with $10M of participating preferred owning 25%, the investor takes $10M off the top plus 25% of the remaining $40M = $20M, versus $12.5M non-participating. Resist participation. At 96% non-participating, you have the market on your side.

Seniority / stacking. Whether later rounds sit ahead of earlier ones (standard) or all preferred shares pari passu. This matters mostly in a bad outcome, which is the outcome you should be modelling.

Anti-dilution. Protects investors if you later issue shares below their price. Broad-based weighted average is market standard and reasonable — it adjusts the conversion price modestly, in proportion to the size of the down round. Full ratchet repraises the entire earlier investment at the new lower price and is punitive; it transfers enormous value from founders and employees in a down round. Reject full ratchet outside a genuine rescue.

Pro rata rights. The right to maintain ownership percentage in future rounds. Standard for major investors. Grant it to leads; be sparing with small holders (see A6).

Dividends. Usually non-cumulative and never paid. Accruing/cumulative dividends (2.4% of Cooley's Q1 2026 deals) function as a hidden increase in the liquidation preference — an 8% cumulative dividend over five years adds 40%+ to what gets paid before common.

The option pool shuffle. The most reliably underestimated economic term. The term sheet says "pre-money valuation $20M, with a 15% post-close option pool." Because the pool is created pre-money, it comes entirely out of existing shareholders — you — not out of the new investor's stake.

Worked comparison on a $5M raise at a $20M pre-money ($25M post):

• No pool: investor 20%, existing holders 80%.
• 15% post-close pool created pre-money: investor still 20%, pool 15%, existing holders 65%.
• 15% pool created post-money (rare, but ask): investor 17%, pool 15%, existing holders 68%.

[Analysis] The pre-money pool is effectively a $3.75M reduction in your valuation disguised as a governance term — in this example, a $20M pre-money with a 15% pre-money pool is economically equivalent to a $16.25M pre-money with no pool. The defence is not to refuse the pool; you need one. The defence is to build a bottom-up hiring plan for the next 18 months, cost it in shares, and negotiate the pool to that number — which is frequently 8–12%, not the 15–20% investors open with. Then explicitly trade: "I'll accept 15% if the pre-money is $23.5M."

Control terms

Board composition. At seed, a 3-person board (two founders, one investor) or often no change at all. At Series A, the near-universal structure is 2 founders / 1 investor / 1 or 2 independents — meaning founders retain control of a 3-person board, or the independent is the swing vote on a 5-person board. The independent director is negotiated jointly and is frequently the most consequential person in the room; take that choice seriously.

Protective provisions. A list of actions requiring preferred-holder consent regardless of board or common vote: selling the company, issuing senior securities, changing the certificate, increasing the option pool, incurring debt above a threshold, paying dividends, changing board size. These are standard and reasonable at market scope. Watch for: consent required for annual budgets, hiring/firing executives, entering new business lines, or setting founder compensation — these convert a shareholder protection into operational control. Also watch for separate series voting, which gives each round an individual veto; a company with four series each holding a veto can be paralysed.

Drag-along. Forces minority holders (including common) to go along with a sale approved by the board and a majority of preferred. Standard, and it means you can be dragged into a sale you dislike.

Founder-specific terms: re-vesting (investors sometimes reset or extend founder vesting at Series A), founder transfer restrictions, and co-sale rights.

No-shop / exclusivity. Typically 30–45 days. Negotiate it down and add a carve-out for the process you already have running. Once you sign, your leverage is gone for that period.

B7. Valuation: how it is actually set, and why a high one can hurt

How it is actually set

Not by DCF. Early-stage valuation is set by:

1. What round size you need, divided by what ownership the investor requires. A seed fund targeting 15% writing a $3M check produces a $20M post-money. This is the actual mechanism for most rounds — the valuation is the output, not the input.
2. Comparables — what similar companies at similar stages raised recently, which is why market data (A8) matters.
3. Competition — the single largest lever a founder controls. Two term sheets moves price more than any argument about TAM.
4. Stage-appropriate heuristics at later stages: ARR multiples, growth-adjusted multiples, Rule of 40.
5. Fund construction — a $50M fund cannot lead a round that requires a $10M check.

409A valuations are a separate thing: an independent appraisal of common stock fair market value for setting option strike prices under US tax law. The 409A comes in well below the preferred price (historically ~20–50% of it at early stages, compressing as the company matures), which is what makes options valuable to employees. It must be refreshed at least annually and after any material event, including a priced round. [CPA/valuation firm required] — do not self-assess.

Why a high valuation can hurt

1. You must grow into it. A $40M post-money seed requires Series A metrics that justify $100M+. If you raised at $40M on a story and reach $800K ARR in 24 months, you are structurally stuck: too expensive for a normal Series A, not good enough for a premium one.
2. It compresses the next round's headroom and raises the probability of a down round, which triggers anti-dilution, damages morale, and makes the following round harder.
3. Employee options become less attractive. A high 409A means high strike prices, meaning less upside per share for the people you are recruiting.
4. It narrows your exit range. A company that raised $60M at a $400M valuation cannot be sold for $150M without the preference stack consuming the common. Founders discover this in year six.
5. It changes who can buy you. Acquirers price off ability to pay and comparable multiples, not your last round.

[Analysis] The reframe that helps: a valuation is not a score. It is a promise about the performance you will deliver before the next raise. Take the highest valuation you are confident you can outgrow within 24 months, and prefer a slightly lower price with a better partner and cleaner terms. A $2M difference in valuation at seed is worth roughly 1–2 percentage points of founder ownership. A bad lead investor on your board for eight years is worth far more than that.

B8. Dilution modelling: a full worked example

A concrete path from incorporation to Series B. Two founders, US C-corp, all numbers rounded.

Step 0 — Formation. Founders A and B split 60/40. 10,000,000 shares. Fully diluted: A 60%, B 40%.

Step 1 — Pre-seed SAFEs (months 4–16). Four post-money SAFEs totalling $2.0M, as in A6, fixing 20.0% of the company.

Step 2 — Seed priced round (month 18). $3M at a $20M post-money, with a 15% post-close pool created pre-money.

Ownership after the seed closes:

The founders raised $5M and hold half the company. (This is close to Carta's observed median of 56.2% collective founder ownership after seed — the example is slightly worse than median because of heavy SAFE stacking.)

Step 3 — Series A (month 40). $12M at a $48M pre-money / $60M post-money → investor takes 20%. The investor requires the pool to be topped back up to 12% post-close, which means issuing roughly 5% of new post-round shares, created pre-money.

Approximate result: everything pre-existing is diluted by ~25% (20% for the new money, ~5% for the pool refresh).

(Carta's observed median after Series A: 36.1%. The example tracks reality closely.)

Step 4 — Series B (month 64). $30M at a $150M pre-money / $180M post → 16.7% for the new investor, plus a ~3% pool refresh, so ~19.7% total dilution.

(Carta's observed median after Series B: 23.0% — the example is better than median, mostly because the Series B dilution here is at the low end. Carta's reported median dilution fell from ~18% to ~16% across seed–Series C during 2025, with Series B at 12.9%.)

What this example teaches

1. Founders are typically below 50% collectively by Series A and below 30% by Series C. This is normal, not a failure.
2. Pre-seed SAFE stacking did more damage than the entire Series B. $2M raised on SAFEs cost 20 points; $30M raised at Series B cost 16.7. Early money is the most expensive money.
3. The option pool is a second, quieter round of dilution at every stage — roughly 3–5 points each time it is refreshed.
4. What matters is the value of your stake, not the percentage. 30% of a $180M company is $54M of paper; 100% of a business worth $3M is $3M. The dilution question is only ever relative to what the capital bought.
5. Model your own version before you sign anything. A spreadsheet with a fully-diluted column, all SAFEs at their caps, and a pool line is 30 minutes of work and is the highest-return 30 minutes in fundraising.

B9. Cap tables

The cap table is the legal record of who owns what. Three rules:

1. One source of truth. Carta, Pulley, AngelList Stack, LTSE Equity, or a rigorously maintained spreadsheet — but exactly one, updated on the day anything changes, reconciled to signed documents. Cap-table drift (the spreadsheet says one thing, the board consents say another) is discovered in diligence and is expensive to fix.
2. Always work in fully diluted terms, including the unissued option pool and every SAFE and note modelled at conversion. "Issued and outstanding" percentages flatter you and mislead you.
3. Keep a side-letter register. Information rights, pro rata, major-investor thresholds, MFN clauses, board observer seats, and ROFRs live in documents outside the cap table and bind you for years.

Common messes: promised-but-never-granted equity; options granted without board approval; advisors with no vesting; a co-founder who left with unvested shares nobody repurchased; the contractor who owns your logo; convertible instruments with inconsistent qualified-financing thresholds. Each is fixable early and painful later.

B10. Board control

Control is not ownership. A founder with 40% can be removed; a founder with 15% and a board majority usually cannot.

Three layers of control:

1. Board of directors. Hires and fires the CEO, approves budgets and financings, approves a sale. This is where operational control lives.
2. Stockholder votes. Elect directors, approve charter amendments and mergers. Typically requires a majority of common and a majority (or supermajority) of preferred voting separately.
3. Protective provisions. A preferred-holder veto that operates regardless of board or stockholder majorities (B6).

Typical evolution: seed — founders control, often 2 founders + 1 investor. Series A — 2 founders + 1 investor + 1 independent (founders retain effective control), or 2 + 2 + 1 with the independent as swing. Series B — founders typically lose a formal board majority. Series C+ — founders are usually a minority of the board.

[Analysis] Practical guidance. Board control is worth fighting for at Series A, and it is worth trading away for a materially better partner or price at Series B — by then, if your board wants you gone, an extra seat rarely saves you. The independent director is the lever founders undervalue: negotiate for a real, jointly-chosen operator rather than a "friendly" name, and understand that the independent's fiduciary duty runs to all shareholders, not to you.

Dual-class structures (super-voting founder shares) are essentially unavailable in private venture rounds outside exceptional leverage, though they are common at IPO.

B11. Investor rights beyond the term sheet

• Information rights. Typically monthly or quarterly financials and an annual budget to "major investors" (holders above a threshold). Set the threshold high enough to keep the list manageable.
• Registration rights. Demand and piggyback rights relating to a future IPO. Almost never exercised as written; negotiate lightly.
• Rights of first refusal and co-sale. The company, then investors, get a right to buy shares a founder or common holder wants to sell; co-sale lets investors sell alongside. These are why founder secondary requires investor cooperation (B15).
• Redemption rights. The right to force the company to buy back shares after a period. Rare and rising: Cooley recorded 6.1% of Q1 2026 deals with redemption, up from 1.8% in Q4 2025. Resist; a redemption right is a put option against your balance sheet.
• Most favoured nation. In a side letter, obliges you to extend better terms given to later investors. Track them.

B12. Use of funds and runway planning

Current norm: raise for 24–30 months. The 18-month rule of thumb is obsolete. [Verified] Carta measured the median gap between funding rounds at 696 days (~23 months) in Q2 2025, up 5% year over year. If the median gap is 23 months and you raise 18 months of runway, you are planning to run out before the median company raises again.

How to size the raise:

1. Define the milestone that unlocks the next round (B2), not the time period.
2. Build a bottom-up plan to reach it: headcount by role by month, then everything else.
3. Add 6 months of buffer for the raise itself, which now routinely takes 4–6 months from first meeting to cash.
4. Sanity-check the dilution: if the resulting raise costs more than ~20–25% of the company, either the milestone is too ambitious or the valuation is too low.

Burn discipline metrics: net burn (cash out minus cash in), runway (cash ÷ net burn), and burn multiple (net burn ÷ net new ARR). Burn multiple under 1.5x is strong at early stage; above 3x is a problem investors will name. Default alive — the state where existing growth and existing cost reach profitability before the money runs out — is the single most useful framing for a board conversation in a tight market.

[Analysis] The trap of raising "just enough." Under-raising is more dangerous than mild over-raising, because a company that misses its milestone with three months of cash left has no leverage and no options. But over-raising has a real cost too: a larger round at a higher valuation sets a higher bar and, at exit, a bigger preference stack. The balance point in the current market is raise for the milestone plus six months, and cut burn early rather than raising a bridge late.

B13. Down rounds and structured rounds

Where the market is

[Verified] Down rounds peaked in the post-2022 reset and have normalized:

• Carta: down rounds were 11.4% of rounds in Q1 2026, described as "back in line with 2019 and 2020 levels" (Carta Q1 2026). Q4 2025 was under 14%, the lowest in three years (Carta Q4 2025).
• Cooley: 11.4% down, 2.6% flat, 86% up in Q1 2026; Q4 2025 was 12.8% down, 7.4% flat, 79.7% up.

For context, down rounds ran above 20% of Carta's book at the 2023–24 trough. [Analysis] The recovery is real but should be read alongside deal count, which is at multi-year lows. Companies that would have taken a down round in 2024 may simply not be raising in 2026 — the improvement in the down-round rate is partly a composition effect, not purely a health signal.

What a down round costs

• Anti-dilution adjustment. Broad-based weighted average adjusts earlier investors' conversion price, issuing them more shares at founders' and employees' expense. Full ratchet is far worse.
• Employee equity underwater. Options struck at the old, higher 409A are worthless at the new price. Expect to need an option repricing or a refresh grant, both of which require board approval and have tax consequences. [Lawyer/CPA required]
• Signal. Customers, candidates, and future investors notice.

Structured rounds — the thing to watch instead of the headline number

A "structured" round preserves a flattering headline valuation by adding economics: multiple liquidation preferences (1.5x–3x), participating preferred, cumulative dividends, full-ratchet anti-dilution, pay-to-play, or redemption rights.

[Verified] These remain minority terms but are measurably present and rising in places: Cooley's Q1 2026 data shows pay-to-play at 7.3% (up from 6.3%), redemption at 6.1% (up from 1.8%), accruing dividends at 2.4%, and 1.8% of deals not at 1x preference.

[Analysis] A structured flat round at a $100M "valuation" with a 2x participating preference is economically worse for common holders than a clean down round at $60M. If you are offered structure, model the exit waterfall at $50M, $150M, and $400M before you compare offers. The number on the press release is not the deal.

Pay-to-play

A provision requiring existing investors to participate in the new round or lose rights — typically conversion of their preferred into common, or into a junior series. Pay-to-play is brutal on investors and, perversely, often good for founders and employees, because it strips preferences from investors who will not fund the company. It appears in rescue financings and recapitalizations.

B14. Bridge rounds

An interim raise between priced rounds, usually on a SAFE or convertible note, usually from existing investors.

Legitimate uses: you are 2–4 months from a milestone that will materially improve terms; a signed-but-not-closed large contract; an acquisition conversation that needs runway to complete.

The honest version: most bridges are raised because the next round is not available yet. Insiders know this. The terms reflect it — bridges commonly price at a discount to the last round, or on a note with a cap at or below the prior valuation, sometimes with a 1.5x–2x conversion multiple or seniority.

[Verified] Carta's analysis of the 2022 seed cohort noted that "bridge rounds are completed at higher rates (both priced and SAFEs), but that's no guarantee of future success" (Carta) — bridges kept companies alive without changing the graduation outcome for most of them.

[Analysis] The decision rule. A bridge is worth taking when you can name the specific metric change it buys and the buyer of the next round. It is worth refusing when it merely postpones the same conversation with less cash and a more complicated cap table. Cutting burn to reach default-alive is often the better answer, and it is the answer an experienced board will respect more.

B15. Secondaries and founder liquidity

The market

[Verified] Secondary activity is at multi-year highs. Carta administered 396 tender offers in 2025, up 62% year over year, with nearly 20% involving Series E+ companies (Carta Q4 2025). In H1 2026 alone, Carta ran 71 tenders worth about $3 billion — the highest H1 in at least six years, with transaction count up 34% and value up 200% year over year (Carta, H1 2026 tender update).

Pricing and structure detail from that report:

• Nearly 70% of tenders were Series C or later. Median offering size: $28.5M at Series C+, $8.5M at seed–Series B.
• Median discount to the most recent primary round price was 0% for tenders at least a year after a round — but the 75th percentile discount rose to 10% in H1 2026, and at least a quarter of tenders included double-digit discounts.
• Median subscription rate 93.1%; median seller participation 57.9%.
• Buyers are growth investors, institutions, hedge funds, and family offices.

Founder secondary

Selling some of your own shares, usually as part of a priced round or a company-run tender.

When it is available: typically Series B and later, and increasingly at Series A for companies with real competitive tension. Common norms: 5–15% of a founder's holdings, or an amount pegged to "life-changing but not life-altering" — enough to remove financial precarity, not enough to reduce motivation.

Why investors allow it: a founder who can pay off student debt and buy a house takes better long-term risks and is less likely to push for a premature $80M exit. Investors say this openly, and it is true.

Why investors limit it: selling a large fraction signals you do not believe. Optics matter with employees, who cannot usually sell.

Mechanics and traps:

• Requires consent. ROFR and co-sale provisions mean your investors must cooperate. It is a negotiation, not a right.
• Tax. Sale of founder common is a taxable event, and QSBS treatment has specific holding-period and issuance requirements. [CPA required] — the difference between qualifying and not qualifying is enormous.
• Price. Founder secondary usually prices at a discount to the preferred round price (because common is worth less), often 20–30%, unless the buyer wants the exposure badly.
• 409A impact. A large common-stock secondary at a high price can raise your 409A valuation, increasing option strike prices for new hires.
• Do it with the employees. A tender that includes employees builds enormous goodwill. A founders-only secondary that leaks builds the opposite.

B16. Exit planning

IPO

[Verified] 2026 is a genuinely open but highly concentrated IPO window. Renaissance Capital counted 106 IPOs priced year-to-date (for deals ≥$50M market cap), down 29.3% from the prior year, but raising $145.8 billion — up 437.9% (Renaissance Capital IPO Stats). Carta counted 34 IPOs priced in Q1 2026 raising $9.9B. Crunchbase reported 32 companies going public above $1B valuations in Q2 2026 alone, and SpaceX's IPO raising approximately $75 billion at a $1.77 trillion valuation — a single deal that distorts every aggregate statistic for the year.

[Analysis] Fewer, much larger. The window is open for scaled, profitable-or-near-profitable companies with predictable growth, and effectively closed for everyone else. Practical thresholds in the current market: roughly $200M+ revenue, 25–40%+ growth, credible path to GAAP profitability, and clean multi-year audited financials. Going public costs $5–15M+ in one-time expense and several million a year in ongoing compliance.

M&A

The realistic path for the overwhelming majority. See A16 and A17 for multiples and mechanics. The key numbers to hold: private SaaS M&A at ~3.1x EV/Revenue in Q1 2026 rising to ~6x above $50M ARR; sub-$10M-EV profitable SaaS at ~3.9x profit with an ~81-day average time to sell.

Acqui-hire

See A17(3). The critical fact bears repeating: consideration is largely retention, not purchase price, and the preference stack usually consumes what is left.

Planning, practically

• Build relationships with potential acquirers years early — through partnerships, not banker introductions. The best acquisitions come from a corp-dev team that already knows you.
• Know your preference stack cold. Maintain a waterfall model showing what common receives at $25M, $75M, $200M, $500M. Update it after every round. Most founders cannot answer "what do I get at $100M?" and should be able to instantly.
• Understand that timing is not yours. Windows open and close on macro conditions.
• Recognize that "no exit" is an exit. A profitable, growing, independent company that pays distributions is a legitimate terminal state — just not one a venture cap table accommodates well, because preferred stock plus a 10-year fund life eventually forces a decision.

B17. Raising versus staying independent: an honest comparison

What the median venture-backed outcome actually is

[Verified] and [Analysis], assembled from the data in this chapter:

• ~17–30% of seed-funded companies raise a Series A within 24 months (Carta). The 2022 cohort managed 17%; historical norm 25–30%.
• Of those that do, a minority reach Series B, and so on. Compounding conservative stage-to-stage rates, roughly 1–3% of seed-funded companies reach an outcome that returns a fund.
• Fewer than 20% of 2017 and 2018 vintage venture funds have returned 1x DPI as of Q1 2026 — eight and nine years in (Carta VC Fund Performance). If the funds mostly have not returned capital, the underlying companies mostly have not exited well.
• Median venture fund IRR for the 2021 and 2022 vintages: 0.5% and 0.1%.
• Median exit for a venture-backed company is a modest acquisition at a price that, after a preference stack, returns capital to investors and comparatively little to common holders. There is no authoritative public median because unsuccessful exits are not reported — which is itself the most important survivorship-bias caveat in this chapter.

The distribution, plainly: a small number of venture-backed founders become very wealthy. A larger number of founders do fine — a $30–80M acquisition after raising $5–15M can produce a genuinely life-changing outcome for founders with meaningful remaining ownership. The largest group spends five to eight years, dilutes to 10–25%, and exits at a price where the preference stack consumes most or all of the common proceeds, or shuts down.

What the median bootstrapped outcome actually is

Weaker data, and the bias runs the other way — there is no Carta for bootstrapped companies, so we only see the survivors and the ones that list themselves for sale.

• Most bootstrapped businesses stay small. The large majority of US businesses never exceed a handful of employees.
• The realistic good outcome is a durable business paying the founder $150K–$1M+ a year, potentially indefinitely, with full control.
• The realistic exit is a profit-multiple sale. Acquire.com's sub-$10M-EV marketplace data: median 3.9x profit, ~81 days on market. A business throwing off $600K a year sells for roughly $2.3M — and the founder keeps nearly all of it, because there is no preference stack and no investor ownership.
• Time to a meaningful outcome is usually longer, and the ceiling is much lower.

The comparison, stated fairly

[Analysis] The genuinely decision-relevant framing: venture capital maximizes expected value while lowering the median outcome. Bootstrapping maximizes the median outcome and the probability of some good result, while capping the upside. Neither is more rational; they encode different risk preferences and different beliefs about the market you are in.

The questions that actually resolve it:

1. Does capital buy speed in this market, and does speed matter? If being first is decisive, raise. If not, do not.
2. What is the honest ceiling? If it is $20M of revenue, venture is a mistake for everyone involved.
3. What outcome would make you happy? If $5M in the bank and control would satisfy you, the venture path asks you to risk that outcome for a shot at a much larger one you will probably not get.
4. Can you tolerate the loss of control? Not the equity — the board seats, the protective provisions, the ability to be fired from the thing you made.
5. What do you believe about your own distribution? Venture is the right choice for the founder who genuinely believes they are building a company worth billions, and an expensive mistake for the founder who says it because it is expected of them.

The hybrid paths are real and underrated: bootstrap to $1M ARR and raise a seed at ten times the price; raise a small angel round and stay capital-efficient; take a modest amount of venture debt or a state SSBCI investment instead of equity; raise once and never again; or take growth equity at $20M ARR with meaningful founder secondary and keep control. Almost nothing in this chapter is a binary.

A note on what this chapter cannot tell you

This is not legal, tax, or investment advice, and nothing here is personalized to your situation. Three areas in particular require professional counsel and are genuinely unsafe to handle from a guide: any securities offering (including friends and family), any non-standard financing instrument, and any tax question touching QSBS, 83(b), option repricing, or founder secondary. The cost of a good startup lawyer for a seed round is a few thousand dollars in advice and $25–60K for a priced round. The cost of a mishandled cap table discovered during Series B diligence is a repriced round or a dead deal.

Finally: the data in this chapter describes a market in an unusual state. Eighty-six percent of US venture dollars went to AI companies in H1 2026, two companies took 43% of global funding, and a single IPO distorted the year's exit statistics. Any benchmark drawn from this period should be treated as a snapshot of an anomaly, not a norm. Re-check every number against the primary source before you bet a plan on it.

Sources

Market data and benchmarks

• PitchBook-NVCA Venture Monitor, Q2 2026 — July 2026
• PitchBook: Q3 2026 — Fewer Deals, Bigger Bets: How Corporate Capital Is Concentrating US AI Venture Activity — 2026
• Carta: State of Private Markets, Q1 2026
• Carta: State of Private Markets, 2025 in Review (Q4 2025)
• Carta: State of Private Markets, Q2 2025
• Carta: State of Pre-Seed, 2025 in Review
• Carta: State of Pre-Seed, Q2 2026
• Carta: VC Fund Performance, Q1 2026
• Carta: VC Fund Performance, Q3 2025
• Carta: Founder Ownership Report 2025
• Carta: Graduation rate from seed to Series A
• Carta: Tender-Offer Activity Reaches a Four-Year High (H1 2026)
• Carta Data Desk index
• Crunchbase News: Global Startup Investment Hit Record $510B In H1 2026 — July 2026

Deal terms

• Cooley: Q1 2026 Venture Financing Report — April 29, 2026
• Cooley: Q4 2025 Venture Financing Report — February 9, 2026

Instruments and accelerators

• Y Combinator: SAFE documents and user guide
• Y Combinator: The YC Deal
• Y Combinator: YC's $500,000 Standard Deal — January 11, 2022
• Techstars: Investment Terms Update — April 17, 2025

Grants and government programs

• Crowell & Moring: SBIR/STTR Programs Reauthorized After Six-Month Lapse — April 2026
• NSF 26-510: SBIR/STTR Phase I, Phase II, Fast-Track solicitation
• ARPA-E: DOE Announces First Projects Under SCALEUP Ready Program in 2026
• European Innovation Council: EIC 2026 Work Programme
• Carta: The State Small Business Credit Initiative (SSBCI) Explained
• NVCA: SSBCI Resources

Equity crowdfunding

• SEC Division of Economic and Risk Analysis: Analysis of Regulation Crowdfunding (PDF) — May 2025, covering May 2016–December 2024
• BusinessWire: FINRA Fines Wefunder $1.4 Million; StartEngine Capital Fined $350,000 — May 2022

Debt and revenue-based financing

• Startup Fundraising: Venture Debt Playbook — Active Lenders & Terms Post-SVB (fundraising-services publisher; terms indicative)
• Pipe: 2026 at Pipe, Building From Strong Foundations (company-published)
• Crunchbase News: Capchase Raises $400M In Debt
• Founderpath: lender comparison pages (vendor-published; self-serving)

Exits, M&A and IPO

• Renaissance Capital: 2026 IPO Market Stats
• Aventis Advisors: SaaS Valuation Multiples 2015–2026
• Acquire.com: Biannual Acquisition Multiples Report, January 2026 (marketplace-published; sub-$10M enterprise value deals)

Process

• DocSend: A (Brief) Anatomy of a Successful Seed Raise — 2019 data (published by a Dropbox-owned document product)

Conflict-of-interest note on sources: Carta, DocSend, Cooley, FE International, Acquire.com, Founderpath and Pipe all have commercial interests adjacent to the data they publish. Carta and Cooley data reflect their own books of business, not the whole market. PitchBook-NVCA, Crunchbase, SEC DERA, NSF, ARPA-E and the European Commission are the closest things to neutral population-level sources used here, and even those have coverage gaps — particularly for companies that never raise institutional capital, which is most of them.