In October 2022, Camden Foley and Samuel Bertain were fired from Session Corp, a cannabis-accessories company they had formed with two close friends. Each founder had started with 25% of the company. After a recapitalization and a later set of stock restriction agreements, the company purported to repurchase Foley's and Bertain's shares for $19.50 each.1

The Delaware Court of Chancery eventually restored the shares. Session's CEO had authority to fire the two men as employees, the court found, while the company's board had never authorized the stock restrictions or repurchases used to cancel their ownership. Foley and Bertain kept their equity because they litigated through a post-trial decision, then recovered discovery expenses after proving serious misconduct during the case.1

The distinction is useful. Litigation was an unusually ugly way to obtain it.

Founder is the word on the podcast introduction. Employee, officer, director, and stockholder are the positions that allocate authority and money. One person can occupy every position on Monday, lose two on Friday, and still tell the literal truth when calling themselves a founder. The title preserves the story. The documents decide what remains.

This article is about the documents signed while a US startup is cheap, friendly, and mostly hypothetical. That is when paperwork feels least urgent and goodwill is doing work it cannot be made to perform later. The governance examples come from Delaware, and the tax rules are federal. Founders elsewhere face equivalent categories with different statutes and remedies, so use the audit and have local advisers verify the result.

"Everything" also needs a definition. It can mean your job and operating control. It can mean the shares, votes, or proceeds that made the risk worth taking. It can reach your personal balance sheet through tax liabilities, guarantees, collateral, and company spending charged to you.

Incorporation creates a boundary between founder and company. The five mistakes below quietly install doors in that boundary.

Mistake 1: Mistaking friendship for governance

Founder vesting is reasonable. It protects the company when someone leaves early with a large stake. A repurchase right is also reasonable. It gives the company a defined way to recover shares that have not vested.

The danger appears when founders treat those provisions as a friendship pact and skip the control map around them. Friendship explains why you trust the other signatures. It does not determine what those signatures authorize.

At Session, three founders kept other jobs while Esther Lenoir Ramirez worked full time. The founders later changed their equity split to give Ramirez and another founder a majority. Foley and Bertain said they accepted the change in reliance on promised governance protections in a founder agreement. That agreement was never executed. The stock restrictions came later, followed by their termination and the attempted $19.50 repurchases.1

The court fixed the unauthorized stock transactions. That remedy should not be confused with a dependable operating procedure. A trial is an expensive place to learn that your employment ended under one allocation of authority, your shares survived under another, and the document carrying your hoped-for protections remained unsigned and disputed.

A second Delaware case shows what happens when the board itself becomes the instrument.

Destiny XYZ began with two equal founders and an agreed five-seat board containing three vacancies. After a 2020 restructuring, Sohail Prasad held about 63.5% of the stock and Samvit Ramadurgam held about 36.5%. Both remained directors. Prasad later appointed two friends to the vacant seats shortly before a board meeting. The reconstituted board approved a 1,850,000-to-1 reverse split followed by a 1-to-1,850,000 forward split, cashing out Ramadurgam and other holders while leaving Prasad as the sole stockholder.2

The Court of Chancery found no fair process or fair price. It imposed a constructive trust that restored Ramadurgam to a 36.5% equity interest and awarded reasonable attorneys' fees.2

A five-seat board with three vacancies sounds unfinished. To the person with appointment power, it can look like inventory.

Before signing founder vesting, a recapitalization, or a voting agreement, make a one-page control map. It should answer:

  • Who can terminate each founder's employment and remove them as an officer?
  • Who can appoint or remove each director, and how are vacancies filled?
  • When can the company repurchase founder shares, at what price, and after which approvals?
  • Who can issue shares, amend the certificate, or change the size of the board?
  • Which directors, stockholders, or classes must approve a financing or sale?

Run the map under two conditions: everyone remains friendly, and the most powerful founder wants you gone. If the second version depends on goodwill, you have described the relationship rather than the governance.

For every answer, write down the document, section, current signatories, and approval threshold. "The founders agreed" identifies a conversation. Your source citation needs the document and section. If a promised protection appears only in an unsigned draft, mark it missing. If two documents appear to conflict, mark the conflict instead of choosing the interpretation you prefer. The purpose is to give counsel a short set of concrete questions while fixing them is still a negotiation among founders.

Run one more check after every board resignation, founder departure, or financing. A control map is a snapshot. A vacant seat, a new class of stock, or a revised voting agreement can change the answer without changing anyone's title on LinkedIn.

Mistake 2: Starting with IP the company does not cleanly own

Seed founders often describe the product as something they built in their spare time. Spare time describes when the work happened. It leaves the ownership question open. The work may still involve a former employer's laptop, repository, data, customer knowledge, confidential information, or invention-assignment agreement.

Anthony Levandowski is the extreme boundary case. He signed employment, confidentiality, invention-assignment, and arbitration agreements while at Google. He left in January 2016, formed Otto, and sold it to Uber in August. Google later obtained an arbitration judgment against him for $179,047,998.64. He filed for Chapter 11 in March 2020 because he could not afford the judgment.3

Precision matters here. The arbitration claims producing that judgment concerned employee poaching. They were separate from Waymo's trade-secret case against Uber and Otto.4 The point for an ordinary founder is narrower: obligations signed at the old company do not evaporate when the new company gets a logo.

The routine version is less cinematic. A contractor wrote the first production code and never assigned it. A founder copied a sales deck containing a former employer's customer material. A critical dependency has a licence nobody reviewed. The company account does not own the domain or repository. Each defect can remain quiet while the founding story does the talking. Then an investor, buyer, enterprise customer, or litigant asks the same question: show me how the company obtained the right to use and sell this.

Build an IP chain-of-title file before the seed data room. Include:

  • signed founder invention-assignment agreements, with any prior inventions or exclusions listed;
  • signed employee and contractor assignments covering the work each person actually produced;
  • a written record of material brought from former employers and the basis for using it;
  • an inventory of material open-source licences and their obligations;
  • proof that the company controls its domains, repositories, datasets, model assets, and customer deliverables; and
  • a short clean-room memo for any feature rebuilt after exposure to a former employer's confidential work.

Have counsel review the file where former-employer language, regulated data, or licence obligations create ambiguity. The memo will be boring. Boring is an excellent property for something that could otherwise become Exhibit A.

Do the review asset by asset. For the core codebase, dataset, model, brand, domain, and first customer deliverables, identify who created the asset, when they created it, which equipment and accounts they used, what agreements bound them at the time, and which signed document moved the rights to the company. "We all worked on it" begins the audit. The next line should name the signed documents that moved the rights.

This also gives you a clean answer during diligence. Instead of asking a buyer to accept the founding story, you can show the signed assignment and the record behind it. The romance survives. The indemnity discussion gets shorter.

Mistake 3: Missing the 30-day tax decision

Restricted founder stock often starts cheap and vests over time, which makes the paperwork feel theoretical. The tax clock treats the transfer as real immediately. Section 83 generally taxes the spread between what a service provider paid and the property's fair market value when the property becomes transferable or is no longer subject to a substantial risk of forfeiture. A timely Section 83(b) election generally moves that income measurement to the transfer date.6

The filing window is short. The current IRS election form says it must be filed no later than 30 days after the property transfer. It also directs the filer to give a copy to the person or company for whom the services were performed.5

Here is a deliberately simplified calculation:

  • Assume a founder buys 1,000,000 restricted shares for $0.001 each.
  • Assume the fair market value on the purchase date is also $0.001 per share.
  • The assumed spread at transfer is $0.
  • Assume 250,000 shares vest one year later when fair market value is $2 per share.
  • Without a valid election, the illustrative compensation spread on that first tranche is 250,000 x ($2.00 - $0.001) = $499,750.

That is taxable compensation in the example, not the tax bill. The calculation excludes later vesting tranches, federal and state rates, capital-gain treatment, and every fact that makes real tax advice expensive. Its purpose is to show why a form that feels irrelevant when the company is worth almost nothing can matter after the first year of progress.

Confirm with a tax adviser whether Section 83 and an election apply to your property and circumstances. If you file, retain the signed election, the stock purchase paperwork, proof of timely delivery to the IRS, and the copy supplied to the company. A calendar reminder proves that you intended to file. It does not prove that you did.

Mistake 4: Taking seed money without mapping control

Founders tend to ask one question after signing a SAFE: how much do I own? It is the answer that fits neatly in a spreadsheet and sounds good in an update.

Keep asking. Ownership is an economic answer at a given moment. Control is distributed across board appointment rights, class votes, stockholder thresholds, protective provisions, information rights, side letters, and approvals required for a sale. Those rights can move while the headline founder percentage still looks reassuring.

Y Combinator's post-money SAFE was designed so the ownership sold by SAFE financing can be measured after all SAFE money and before the new money in a priced round. YC also publishes an optional pro rata side letter.7 Those documents help model a specific instrument. They do not reveal every right promised across every email, side letter, or later financing document.

The National Venture Capital Association's model set makes the document sprawl visible. A priced round can use a charter, stock purchase agreement, investors' rights agreement, voting agreement, and right-of-first-refusal and co-sale agreement. "Model documents" is a tidy label for rights that only make sense when the whole stack is read together. Each model is a starting point with options.8

Maintain two models after every signature:

  1. The economics model. Show fully diluted ownership after all SAFEs convert, option-pool changes, and the proposed round. Then model proceeds across several exit values. Our dilution calculator helps with the ownership math; how founders exit with nothing covers the liquidation-preference and waterfall layer.
  2. The control model. Show who appoints directors, which matters require board or class approval, what each side letter promises, and which holders can approve or block a financing or sale.

If the cap table says founders own 72%, the screenshot showing 72% will look excellent. It does not tell you whether one founder can lose their board seat, whether an investor has a contractual consent, or whether the votes needed for a sale sit somewhere else.

Stress-test the control model instead of admiring the current-state diagram. Ask what approvals are needed if a founder is absent, if the board disagrees, if a new lead investor demands a seat, and if a low-priced acquisition arrives before the next round. Beside each answer, show who can act alone and who can block. A right that appears harmless while everyone wants the same financing can become decisive when the preferred outcome changes.

Read founder equity dilution for the ownership model and angels who leave you alone for an investor-rights screen. Then ask counsel to reconcile both models against the signed documents. Founder confidence tends to live in the spreadsheet. Counterparty rights live in the consent matrix. Both should describe the same company.

Mistake 5: Letting company debt follow the founder home

The company borrows. The founder signs. The lender may describe the second signature as support. The documents determine whether that support reaches the founder's assets.

A personal guarantee gives a creditor recourse against the guarantor according to its terms. The possible shapes are visible in the US Small Business Administration's own guarantee instructions: unlimited liability, a maximum dollar amount, a percentage, a time limit, specified collateral, or other recourse conditions.9 Startup lenders and landlords use their own documents, so inspect the actual obligation in front of you.

Rand Fishkin has described the ordinary early-company path. He said company borrowing across banks, lenders, and credit cards grew to roughly $450,000 to $500,000. After the balance passed a point where minimum payments became unmanageable, interest and penalties accumulated. Settlements later damaged his credit. Fishkin said he avoided bankruptcy because he believed family assets, including his grandmother's house, could be exposed.10

It did not begin as a plan to endanger a family home. It began as money for lean months.

At the extreme end sits Faraday Future founder Yueting Jia. In a statement about his 2019 Chapter 11 case, Jia's debt project team said more than 90% of his debts came from business guarantees. The team said his plan transferred all personal assets recognized by the US court, including his economic interest in Faraday Future, to a creditor trust, and put his unpaid net debt at about $2 billion.11 Reuters reported in July 2020 that Jia completed the personal bankruptcy process and also described his personal net debt as roughly $2 billion.12

Jia's was a late-stage collapse at enormous scale. It is not evidence that he signed those guarantees during a seed round. It shows the endpoint of the same contractual mechanism: company obligations become personal obligations because somebody agreed that they would.

Create a guarantee register covering every loan, lease, card, equipment facility, and other company obligation for which a founder may be personally liable. For each one, record:

  • the creditor and underlying company obligation;
  • the founder or other person who signed;
  • the maximum personal exposure, if there is one;
  • the term, expiry, and any burn-off condition;
  • pledged collateral and the scope of recourse;
  • the events that trigger liability;
  • release mechanics after repayment, refinancing, or a financing milestone; and
  • whether the guarantee is unlimited, capped, percentage-based, or limited to specified acts.

Then negotiate from the register. Ask whether an unlimited guarantee can become capped, whether the cap can decline with the principal, whether recourse can expire after a clean payment period, and whether a financing can trigger release. The creditor may call its request standard and may decline. You will still have converted a vague feeling of support into a visible personal risk decision.

Add the exposures together under a bad month. The lease, card balance, bridge, and equipment facility may have been approved at different times by different people. Your personal balance sheet experiences them at once. Record whether one default can affect another obligation, whether the same collateral supports several debts, and whether a release has to be requested in writing after its condition is met. Record those answers from the executed documents in the guarantee register and confirm them with your lawyer.

Also record founder money already inside the company. Mark each transfer as salary left unpaid, an expense awaiting reimbursement, a documented loan, or an equity contribution based on the advice you receive. "I paid it personally because the card was handy" describes the transaction. It does not classify it.

Build the founder downside file

These mistakes share a timing problem. Founders call the document administrative when the company has little value. A counterparty calls it decisive after a termination, vesting date, diligence request, financing, sale, or default gives someone a reason to enforce it.

Create one founder downside file before the next signature. It should contain:

  • founder stock purchase agreements, vesting schedules, repurchase terms, and Section 83(b) evidence;
  • the certificate, bylaws, board appointments, voting agreements, side letters, and a current consent map;
  • the cap table, SAFE conversion model, and downside exit waterfall;
  • founder, employee, and contractor IP assignments, plus former-employer and open-source reviews;
  • the guarantee register and a record of every personal source of company funding;
  • the employment and officer documents for each founder; and
  • the lawyer and tax adviser responsible for confirming each unresolved item.

Do this before the next financing, lease, corporate card, or bridge loan. Treat the file as the company record with the optimism removed. Afterwards, the same audit becomes a negotiation with someone who already owns the leverage.

Before you sign, trace how each company document or transaction can affect your job, equity, control, tax position, or personal assets. Put the governing document, approval threshold, deadline, and maximum exposure beside each one.


Sources
  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12