Incorporator The Journal of U.S. Formation
Start Free Quiz
Editorial 8 MIN READ

How to split equity at formation

Authorized shares, restricted stock, the new IRS Form 15620, and the cap-table decisions that survive every future round

The Editors
Staff, Incorporator.org
8 min read Updated
Contents 8 sections
  1. Authorized shares and par value
  2. Issuing founder stock at a low basis
  3. The 83(b) election, thirty days, and the new Form 15620
  4. The split itself: equal, role-weighted, or dynamic
  5. Vesting, cliffs, and acceleration
  6. The option pool and the advisor shares
  7. The QSBS clock runs from the issuance, not the raise
  8. Sources

ounder equity at formation is a set of defaults that almost nobody revisits once they calcify. Get the first two weeks right and the cap table carries you through a decade; get them wrong and you pay a tax lawyer to untangle the damage at the worst possible moment. This guide is for a two or three-person team incorporating a Delaware C-corp in the spring of 2025. If you are still weighing an LLC rather than a corporation, the Delaware formation guide covers that side of the choice.

Authorized shares and par value

The first number on the Certificate of Incorporation is total authorized shares. The market-standard default is 10,000,000 shares of common stock with a par value of $0.00001 per share. That figure is not arbitrary; it is engineered to keep Delaware's franchise tax at the minimum.

Delaware calculates the annual franchise tax under one of two methods. The default method printed on the state's invoice is the Authorized Shares Method, which charges by raw share count; a corporation with 10 million authorized shares receives a bill well into five figures under that calculation. The alternative, the Assumed Par Value Capital Method under 8 Del. C. § 503(a)(2), computes tax based on gross assets and issued shares, and for a typical early-stage company drops the bill to the $400 minimum plus the $50 annual report fee. A very low par value is what makes that second calculation come out small. Authorize the same 10 million shares at $0.01 par and the math no longer rescues you.

None of this is a tax dodge. It is the statute doing exactly what the statute says. It is also why every YC-standard incorporation uses the same two numbers, and why you should not invent your own.

Issuing founder stock at a low basis

You issue founder stock almost immediately after incorporation, usually within days. The mechanism is a Restricted Stock Purchase Agreement, not an option grant. The founders are buying the shares outright, at a price equal to the then-current fair market value, which on day one is par or just above par. The canonical figure is $0.0001 per share, ten times par, so a founder receiving 4,000,000 shares writes the company a check (or contributes IP of equivalent value) for $400.

Paying something establishes basis and makes the purchase real. Paying almost nothing is defensible because the shares are not yet worth anything; there is no operating company, no customers, no revenue, nothing the IRS could reasonably mark up. Delay issuance until after a seed round and the fair market value has moved, which turns the same founder stock into compensation income at grant and produces a W-2 event with no liquidity to pay the tax. The fix for that problem is expensive; the fix for the original problem is to do the issuance on day one.

Founder shares are almost always "restricted" in the Section 83 sense. The founder owns them, but the company holds a right to repurchase the unvested portion at the original purchase price if the founder leaves. This is sometimes called reverse vesting, because the shares are fully issued up front and vesting is the schedule on which the repurchase right lapses. The tax treatment is different from an option grant, and the difference is why the 83(b) election exists.

The 83(b) election, thirty days, and the new Form 15620

Internal Revenue Code § 83(b) lets a recipient of restricted property elect to be taxed at transfer rather than at vesting. For founder stock, the spread between purchase price and fair market value at issuance is essentially zero, so the election is into a zero-dollar tax event. Without it, each future vesting tranche is a taxable event at that quarter's fair market value, which for a company that raises a Series A can mean ordinary income on paper gains the founder cannot sell to cover the bill.

The election must be postmarked within 30 days of the stock transfer. The 30-day clock is in Treas. Reg. §1.83-2(c), and it has no extension, no cure, and no reasonable-cause relief. Miss it and the election is gone.

Until late 2024, the election was a one-page letter drafted from a template in Rev. Proc. 2012-29. In November 2024 the IRS released Form 15620, the first official agency form for 83(b) elections. The form is voluntary, not mandatory; a taxpayer can still file a properly drafted letter under §1.83-2. But the form exists, it is numbered, it has a catalog number (95376D), and investors, accountants, and the service centers themselves now treat it as the default. File Form 15620, send it by certified mail to the IRS service center where you file your personal return, keep the green card, and send a copy to the company. The statute puts the filing obligation on the taxpayer, not on counsel and not on the company.

Electronic filing of Form 15620 became available on the IRS site in 2025. The paper route still works, and for a formation event where the 30-day window is the only thing that matters, the belt-and-braces approach is to file electronically and retain a paper backup.

The split itself: equal, role-weighted, or dynamic

Once the mechanics are done, the actual split is a social decision with downstream tax and governance consequences.

The simplest split is equal. Two founders, 50/50. Three founders, a third each. Equal splits signal partnership, minimize early negotiation friction, and are defensible to any future investor. They also leave no tiebreaker when the founders disagree about whether to take the acquisition offer or keep building, which is how a surprisingly large number of small companies die. If the split is equal, write the deadlock mechanism into the stockholders' agreement: a coin flip, a mutual buyout at a formula price, or an outside board seat with a deciding vote.

Role-weighted splits assign more equity to the person carrying the heavier load, typically the CEO, the founder who quit their job first, or the founder who contributed the underlying IP. A 60/40 or 55/45 split is more common than the folklore of equal splits suggests. The weakness is that the weights are set on day one based on day-one facts, and eighteen months later the quieter cofounder has written most of the code while the CEO was raising. Vesting softens this; it does not fix it.

Time-weighted or dynamic equity models, the best-known being Mike Moyer's Slicing Pie framework, try to make the split reflect actual contribution over time. Every hour worked and every dollar invested moves the allocation. The model is hard to run once there are outside investors; institutional capital wants a fixed cap table it can price, not a live spreadsheet. Dynamic equity is a reasonable bridge before the first priced round and a poor structure after it.

Vesting, cliffs, and acceleration

The market default for founder and employee stock is a four-year vest with a one-year cliff. Nothing vests in the first twelve months; at the twelve-month mark, 25% vests in a single step; the remaining 75% vests monthly over the following thirty-six months. If a founder leaves before the one-year cliff, the company repurchases all of their shares at the original purchase price. After the cliff, the company repurchases only the unvested portion.

Two acceleration clauses come up in every term sheet. Single-trigger means some or all unvested stock vests automatically on a change of control. Double-trigger means it vests only if the founder is terminated without cause (or resigns for good reason) within some window after a change of control, typically twelve months. Acquirers prefer double-trigger; founders prefer single-trigger. The market answer is usually double-trigger with a carve-out, and that is what the first institutional term sheet will ask for.

The option pool and the advisor shares

The last two pieces of a formation cap table are the employee option pool and any advisor grants.

Institutional seed and Series A term sheets almost always require a fully diluted option pool of 10% to 20% before the round closes, which means the pool comes out of the founders' shares rather than out of the new investors'. This is the pre-money pool shuffle, and it is the single largest hidden dilution founders face at their first priced round. Pre-setting a pool at formation does not eliminate it; investors still demand fresh room. It does force the founders to think about how much of the company is reserved for future hires before anyone has been hired.

Advisor shares are typically granted under the Founder/Advisor Standard Template (FAST) agreement maintained by the Founder Institute, which codifies a grid running from roughly 0.1% up to about 1.0% of common stock, depending on the company's stage and the advisor's engagement level, vesting monthly over two years with no cliff. FAST is a convention. Using the convention saves hours of bespoke negotiation per advisor and produces a cap table later investors recognize on sight. Large deviations from the FAST grid read as a flag.

The QSBS clock runs from the issuance, not the raise

The piece that founders in 2025 most often underweight is IRC § 1202, the qualified small business stock exclusion. For stock issued by a domestic C-corp that meets the § 1202 requirements and is held for more than five years, a non-corporate holder can exclude the greater of $10 million per issuer, or ten times the aggregate adjusted basis of the stock, from federal capital gain. The gross-assets test at issuance is $50 million. The exclusion is 100% for stock issued after September 27, 2010, which covers essentially every active founder today.

The clock starts on original issuance. Founder stock issued on incorporation day begins accruing the five-year holding period immediately, which is the main tax-design reason to issue shares promptly and pay for them. Shares that come out of the option pool later begin their own five-year clocks on their own grant dates. Stock that converts from SAFE or convertible-note instruments starts its clock on the conversion date, not on the date the instrument was signed. A founder who delays issuance by twelve months delays the QSBS eligibility date by the same twelve months and, in a fast-moving company, can push a five-year exit into a four-year exit that does not qualify.

Two planning points. First, keep the company's gross assets under $50 million until after all founder and pool stock has been issued, because once assets cross that threshold the § 1202 clock stops ticking for new issuances. Second, the QSBS treatment is an issuer-level determination; an S-corp or LLC does not qualify, which is another reason the C-corp default is the C-corp default.

The rule of thumb: split the founder equity you are willing to defend in year three, file Form 15620 in week one, and remember that the QSBS clock starts on the day the stock is issued.

Sources

Keep reading

More from the journal.