Leather-bound corporate minute book

The first company I helped build didn't start with a product. It started with a filing to the Delaware Division of Corporations and a wire transfer of $139 that I expensed to nobody, because there was nobody to expense it to yet. That part was easy. The part that cost me was everything I did in the wrong order afterward — stock issued before vesting existed, an 83(b) election I almost mailed on day 29, a cap table that lived in a spreadsheet with a tab named "final_v3." So this is the walkthrough I wish I'd had: the actual sequence, the actual numbers, and the actual reasons investors will care about paperwork you think is boring.

One framing note before the steps. Incorporation is not a legal chore you finish so you can start the real work. It is the foundation every future negotiation sits on. When a fund diligences you, they are not reading your charter because they love charters. They are looking for leverage — a missed filing, a share count that doesn't reconcile, a founder who never signed anything. Every gap becomes a reason to retrade the price. File clean, and you take that leverage off the table before the first meeting.

You can incorporate anywhere. Your accountant's cousin will suggest Wyoming because the fees are low. Someone on a forum will swear by Nevada. And every venture fund you will ever pitch expects Delaware, so the Wyoming advice costs you a conversion later, at lawyer rates, at the worst possible time — which is to say, during your round.

Three real reasons, none of them mystical. First, the Court of Chancery. Delaware decides corporate disputes in a dedicated court with judges — chancellors — who hear nothing but business cases, and there are no juries. If you ever end up in a fight over your board, your stock, or your sale process, you get a written opinion from a specialist citing two centuries of precedent, not a coin flip in front of twelve people who were promised parking validation. Predictability is the product Delaware sells, and it sells it to both sides of the table, which is exactly why both sides accept it.

Second, the law itself. The Delaware General Corporation Law is the most-litigated, most-interpreted corporate statute in the country. Your lawyer can answer almost any governance question — can the board do this, what happens if a director resigns mid-dispute, what vote do we need to sell — without guessing. When the answer to a question is a known case instead of an open one, deals close faster and fights settle cheaper.

Third, and this is the one founders underrate: investor familiarity is a tax you either pay once or pay forever. Every VC partnership agreement, every fund's outside counsel, every diligence checklist assumes a Delaware C-corp. Not an LLC — funds can't hold LLC interests cleanly because of pass-through tax problems for their LPs. Not an S-corp — S-corps can't have more than 100 shareholders or any non-individual shareholders, so the fund itself would blow up your election. The C-corp is the only structure that scales to preferred stock, option pools, and an exit. Converting later costs $10,000 to $40,000 and adds two months to a closing. Or you can spend an afternoon now. Pick the afternoon.

The certificate of incorporation — the charter — is a short document you file with the Delaware Division of Corporations. You can file it yourself online, by fax if you enjoy nostalgia, or through a registered agent's filing service. The state fee is $89 for the standard filing. Want it back in 24 hours instead of a week? Add $50 for the expedite. Same-day service exists for more, and it's rarely worth it unless a wire is literally waiting on the filing, which happens more often than you'd think the week before a close.

Keep the charter minimal. This is the mistake the over-eager founders make — they draft a beautiful charter full of bespoke provisions, and every bespoke provision is something a future investor's counsel will flag, and something you'll pay to amend out. Name, registered agent, purpose clause (keep it broad — "any lawful act"), the stock authorization, the incorporator's name, and the standard indemnification and Section 102(b)(7) exculpation language that shields directors from personal liability for duty-of-care claims. That last clause is non-negotiable. No serious director joins a board without it, and your future independent directors will check.

On share counts, here's the move most guides skip: you don't need to authorize ten million shares on day one. Authorize 300 shares of common stock at $0.0001 par value. Why so few? Delaware franchise tax. Under the authorized shares method, a corporation with 5,000 or fewer authorized shares pays the $175 annual minimum. Authorize ten million shares on day one and the default calculation can spit out a bill in the thousands unless you use the assumed par value capital method and file the math correctly — which founders routinely don't, and then panic-pay a $7,000 tax bill on a company with $600 in the bank. Start small, amend to millions of authorized shares right before your priced round, when you'll need room for preferred stock anyway and when investors' counsel will effectively proofread the amendment for you.

The incorporator — often your lawyer's paralegal, sometimes you — files the charter, then signs a short action appointing the initial board and resigning. That piece of paper matters more than it looks. It's the first entry in your minute book, the leather-bound thing in the photo above that every diligence team will eventually ask to see and that every founder assumes is decorative. It is not decorative. It is where the state of Delaware, conceptually, believes your company lives.

Delaware requires every corporation to maintain a registered agent with a physical address in the state. The agent's job is to accept service of process — lawsuits, subpoenas, official state notices — and forward them to you. That's it. The market rate is about $100 a year; the range runs from roughly $50 to $300, and the expensive ones are not three times better at forwarding mail. Shop on price. It's a commodity service wrapped in a suit.

Unless you personally live in Delaware and enjoy receiving summonses at home, you cannot be your own agent, and you don't want to be. The failure mode here is ugly: the agent resigns for non-payment, the state sends notices to a dead address, someone sues your company, and a default judgment lands before you even know there's a case. I watched a founder discover a two-year-old default judgment during diligence for his seed round. The round survived. The renegotiation of it was not pleasant, and the whole mess traced back to a $99 invoice nobody paid. Set the agent's renewal on autopay and never think about it again, which is the correct amount of thinking.

The charter goes to the state. The bylaws stay home — they're the internal operating manual: how directors get elected, how meetings get called, what officers exist, how many votes pass a resolution. The board adopts the bylaws, and here's the part that surprises people: your initial board can be a single founder. Delaware lets one person be the sole director, and that same person can hold every officer title at once — president, secretary, treasurer, the whole row of hats. Two co-founders? Both can be directors, and two out of two is a quorum. You do not need a third person, a board observer, or anyone's permission.

Run the early company by written consent instead of formal meetings. A unanimous written consent of the board is a one-page document that approves a list of actions — adopting bylaws, appointing officers, authorizing the stock issuances, approving the form of the invention assignment agreement, accepting the IP assignment from the founders. Every one of those approvals needs to exist on paper before you issue a single share, because the stock purchase agreements that follow reference board approval, and diligence will check the dates in order. Consent first, purchase second. Reverse the order and a careful lawyer — the kind your acquirer will employ — will make you ratify the whole chain retroactively, and every retroactive fix invites the question of what else you did backwards.

One obligation founders forget: Delaware requires an annual meeting of stockholders, or an annual written consent in lieu of one, to elect directors. Miss it for three years and the DGCL still lets anyone petition the Court of Chancery to force one. It takes twenty minutes to sign a consent re-electing yourselves each year. Put it in the minute book next to the incorporator's resignation. Future you, sitting in diligence, will be weirdly grateful.

Founders don't "get" their stock. They buy it, and the distinction matters to the IRS. Each founder signs a stock purchase agreement and pays the company for the shares — cash is cleanest, and assigning intellectual property to the company counts as consideration too, which conveniently doubles as the founder IP assignment the next guide covers. With 300 authorized shares and a $0.0001 par value, the math is almost comical: issue 120 shares to each of two co-founders, keep 60 in reserve, and each founder pays $0.012. I have literally taped a penny and a fifth of a nickel's worth of sincerity into a minute book. Pay something, though. Shares issued for nothing invite the argument that they were never validly issued at all.

Now the clock. If your founder stock vests — and it should, keep reading — then what you bought is restricted stock, and the IRS default rule for restricted stock is vicious: you pay ordinary income tax on each tranche as it vests, valued at whatever the company is worth on that vesting date. Company worth $20 million at your two-year vesting mark? Congratulations, you owe tax on your vested slice of $20 million in paper you can't sell. Founders have taken out loans to pay tax on stock in companies that later died.

The escape is the 83(b) election: a one-page filing that tells the IRS to tax everything now, at today's value — which, at formation, is the par value you just paid, so the tax is effectively zero and your long-term capital gains clock starts immediately. The deadline is 30 days from the purchase date. Not 31. Not "about a month." Thirty, and there is no extension and almost no cure. Mail it to the IRS by certified mail, keep the receipt, keep a copy with proof of mailing in the minute book, and mail a copy to the company. The IRS published a standard form for it — Form 15620 — in late 2024, so the "my lawyer drafted a bespoke version" excuse is gone. Set three calendar reminders the day you sign the purchase agreement. This single piece of paper has a higher dollars-per-word ratio than anything else you will ever sign.

Founders hear "vesting" as investors taking something. Reframe it. Standard founder vesting is four years with a one-year cliff: nothing vests for twelve months, a quarter vests at the cliff, and the rest vests monthly after that. The cliff exists for one scenario — your co-founder leaves in month seven. Without vesting, they walk away owning a third of the company forever, and every future investor prices that dead equity against you, the person still showing up. With vesting, they leave with nothing before the cliff and a sliver after it. Vesting is founders protecting each other. Investors just get to free-ride on the protection, and honestly, fine.

Adopt it voluntarily on day one. If you wait until a term sheet forces it on you, the investor's version will restart your clock from the closing date — meaning two years of sweat get un-vested as a condition of funding. I've seen exactly that ask, presented as standard, at 11 p.m. before a signing. The founder who already had four-year vesting running from incorporation got credit for time served. The one who didn't, didn't.

Acceleration is the sub-clause worth fighting over. Single trigger acceleration means your unvested shares vest automatically on an acquisition. Sounds great; investors hate it, because acquirers discount any deal where the key people can vest out and walk the day after closing, so single trigger can literally lower your exit price. Double trigger acceleration needs two events: the company gets acquired and you get terminated without cause (or pushed out via a constructive dismissal definition — read that definition twice) within some window, usually 12 months. It's the founder-friendly version that survives diligence, and it protects you from the acquirer who buys the company and fires the founders to recapture the unvested shares. But it's still less common than plain no-acceleration vesting, so you usually have to ask for it, and the best time to ask is when you set the vesting schedule — before there's an investor at the table to say no. Ask for double trigger on all founder stock and on any options you grant yourself. Worst case, they negotiate it down later. The clause you never papered can't be negotiated at all.

The moment you want to grant stock options to an employee, you need to know what your common stock is worth — because the option's strike price must be at or above fair market value on the grant date, or Section 409A of the tax code turns the option into a deferred compensation disaster. The penalties land on the employee, not you: immediate income recognition, a 20% penalty tax, plus interest. Nothing ends a recruiting conversation like "our options might be a tax trap."

So you hire an independent valuation firm to produce a 409A report. Early stage, that runs $2,000 to $5,000, and the cap table platforms bundle it cheaper. The report gives you a defensible FMV — often a fraction of the preferred price, since common lacks the preference stack — and as long as you granted at or above the report's number, you get safe harbor: the IRS has to prove the value was wrong, rather than you proving it was right.

Two rules to calendar. A 409A is good for 12 months maximum. And any material event — closing a funding round, signing a term sheet, a revenue jump that changes the story — voids it early. Grant options off a stale valuation and the safe harbor evaporates. The rhythm is simple: new round closes, new 409A, then new grants. Founders who batch their option grants right after each financing pay less, sleep better, and never have the conversation where an employee's accountant calls the company a liability.

Incorporate in Delaware and you're domestic in exactly one state. Everywhere else you actually operate — an office, employees, in some states just enough repeated business — you're a "foreign" corporation, and that state wants you to register. This is foreign qualification, and skipping it is one of those savings that compounds into a bill.

The usual suspects: California wants a foreign qualification filing of about $100, a Statement of Information around $25, and then — the famous one — a minimum franchise tax of $800 a year, due whether or not you make a dime. New York runs about $225 for the application for authority plus a comically small $9 biennial statement. Texas charges $750 to qualify and then generally leaves you alone under its no-tax-due franchise threshold, which covers any company you're likely to run pre-Series A. Penalties for operating unqualified vary — back fees, fines, and the sharpest one: in many states you can't maintain a lawsuit in that state's courts until you qualify. A founder who can't sue to collect a six-figure receivable because she saved $800 is a founder who learned compounding the hard way.

And remote employees count. Hire an engineer who works from Austin and congratulations, you may now be transacting business in Texas. Don't panic-register in all fifty states — but when you hire someone in a new state, add "check qualification there" to the same afternoon you run payroll setup. The fees are trivia. The retroactive cleanup is not.

The EIN is your company's Social Security number, and the IRS gives it away free on its website in about ten minutes if the responsible party has a US taxpayer ID. You need it for the bank account, for payroll, for tax filings, for basically everything. No SSN or ITIN — common for foreign founders — and you file Form SS-4 by fax and wait about four weeks, which is a fine reason to file it the same week you incorporate. Whatever you do, don't pay a formation service $200 to obtain a free number. They are billing you for their typing.

Banking deserves a post-2023 update. For a decade the default answer was Silicon Valley Bank — they understood startups, they moved fast, and then in March 2023 they were gone in a weekend, and every founder in America learned the phrase "uninsured deposits" at the same time. SVB's corpse got sold, the brand still exists under new ownership, and the ecosystem moved on: Mercury and Brex picked up enormous startup share with fast onboarding and sane interfaces, and First Republic — which itself followed SVB into FDIC receivership two months later and now operates inside JPMorgan — is back to courting startups with the confidence of an institution that hopes you have a short memory. The durable lesson isn't which bank. It's that you want two of them. Split your cash, keep payroll able to run from either account, and never again let one institution's bad bond portfolio stand between you and making payroll.

Opening the account, bring the stack: EIN confirmation letter, filed charter, bylaws, the board consent authorizing the account, ID for every signer, and usually a cap table. Banks ask for the cap table because regulators make them care who owns what. Which brings us to the document you'll update more than any other.

From the first founder share, maintain one authoritative cap table: every share issued, to whom, on what date, at what price, under what agreement, with what vesting. A spreadsheet is genuinely fine at the start — a good spreadsheet beats a sloppy platform account. Columns: stockholder, security type, shares, price paid, date, vesting schedule, and the document reference. Every entry should trace to a signed piece of paper in the minute book.

Switch to a real platform — Carta or Pulley are the two everyone uses — at around ten stakeholders, or when you start granting options, whichever comes first. Options are what break spreadsheets: exercises, terminations, expirations, 409A-linked strike prices, and the ever-shifting fully-diluted denominator. The platforms also run the waterfall math — who gets what at various exit prices — which is the exact arithmetic investors will run against you in a term sheet negotiation, and which you should therefore be able to run against yourself. Carta is the incumbent with the ecosystem and the pricing to match; Pulley is the cheaper, founder-loved upstart. Either beats the founder who shows up to diligence with "cap table final_v3.xlsx" and a nervous laugh.

Because here is what diligence does with your cap table: they rebuild it from the documents and compare it to yours. Every mismatch is a question, every question is a week, and every week is leverage. A cap table that reconciles to the paper on the first pass tells the other side your house is in order — and quietly removes the excuse for the "we found some issues" price conversation.

Everything above, compressed: file the charter in Delaware ($89, plus $50 if you're impatient); hire the registered agent (~$100 a year, autopay); incorporator appoints the board; board adopts bylaws and approves the stock issuances by written consent; founders buy their restricted stock for par value and assign their IP; everyone mails an 83(b) within 30 days with certified-mail proof; get the EIN the same week; open two bank accounts; start the cap table ledger; foreign-qualify wherever you hire; and get a 409A before the first option grant, refreshing it every 12 months or after any material event. Total cash outlay at formation, excluding lawyers: a few hundred dollars. Total cost of doing it in the wrong order: a repriced round, a tax bill on phantom income, or a co-founder-shaped hole in your ownership forever.

The next document in the stack is the one diligence reads second, right after the charter: the IP assignment. The code has to belong to the company — not to you, not to your co-founder, not to the contractor who built the prototype in a weekend. Our IP assignment guide covers that paper in the same detail, and the SAFE guide picks up the story at your first real check. Do the boring parts now. Boring, done early, is the cheapest alpha in this entire game.

Venture Legal Hub is not a law firm. This content is for informational purposes only and does not constitute legal advice.