On this page
Ask a founder to show you their options and they open the plan. Ask for their shares and they open the register. Ask for their warrants and, more often than not, there is a pause, a scroll through email, and eventually a PDF attached to a loan agreement from three years ago. The warrant is real, it dilutes everyone, and it lives nowhere that the company treats as a record.
This is a field-notes piece, not a legal guide. Warrants are the single equity instrument we most often find missing from a set of records that is otherwise well kept, and the reason is structural rather than careless. This is what a warrant is, why it slips out of the record, the four ways that absence surfaces in diligence, and what the file would have to look like for the warrant to be counted like every other claim on the company.
What a warrant actually is
A warrant is a contractual right to buy a set number of shares, at a set price, for a set period of time. In that shape it looks a lot like a stock option, and the confusion between the two is part of why warrants get lost. The difference that matters for your records is who holds them and why. Options go to employees, directors, and advisors under a plan the board adopted. Warrants go to counterparties in a transaction: a venture-debt lender who wants equity upside alongside the loan, a bank arranging a facility, a landlord, a strategic partner, a placement agent taking part of its fee in coverage, or an investor sweetening a bridge note.
Because a warrant is issued as a term of some other deal, its paperwork is drafted by the other side's counsel and buried in the agreement that governs that deal. It is not created inside your equity system, it does not sit under your option plan's pool, and it does not appear on the share register because no shares have been issued yet. It is a promise about shares that will exist only if and when the holder exercises. Until then it is invisible to every artifact your company treats as the source of truth, and fully visible only to the person on the other side who is holding it as an asset.
Why records forget warrants
Options break because their dates drift across the plan, the grant, and the vesting start, a dynamic we walked through in where stock options live in your records. Warrants break for a simpler reason: nobody owns the record. The lender's counsel drafts the warrant. The founder signs it as one of a stack of closing documents for the loan. The cap table is maintained by the CFO or a founder who may never see the loan file. The minute book is maintained by outside counsel who papered the corporate approvals but not necessarily the commercial terms. Four people touch the transaction and not one of them treats the warrant as a line item they are responsible for carrying forward.
The same gap opens with SAFEs and convertible notes, which is why we treat them as records the moment they are signed rather than at conversion, covered in where SAFEs and convertible notes belong in your records. A warrant is worse in one respect: a SAFE round is usually a deliberate financing event that someone is tracking, while a warrant is frequently a rider on a deal that was about something else entirely, a loan or a lease or a fee. The equity consequence is a side effect of a non-equity transaction, and side effects are exactly what records forget.
The warrant that never reached the cap table
The most common finding is the plainest one: the fully-diluted cap table the company hands to a buyer or a new investor does not include the warrants. The founder is not hiding anything. The warrant was signed, filed with the loan documents, and never entered anywhere the cap table draws from, so the cap table is complete as far as its maintainer knows and wrong as far as the company's actual obligations go.
What it costs at diligence: the buyer's counsel reads the debt file, finds a warrant for, say, one percent of the company on a fully-diluted basis, and asks why it is not on the cap table. Now every number the seller produced is suspect, because the one document that would have caught the error, the fully-diluted table, is the document that was wrong. The reviewer stops taking the cap table at face value and starts reconstructing it from primary documents, which is slower, more expensive, and paid for by the seller. A one-percent warrant is not a large economic item. Discovering it in the data room rather than in the disclosure schedule is what does the damage, because it converts a clean cap table into one that has to be proven.
Coverage that moves after issue
Some warrants are for a fixed number of shares. Many are not. Venture-debt warrants are frequently written as coverage: a percentage of the loan amount, or a percentage of the company, that translates into a share number only at a later reference point such as the next priced round. Placement-agent warrants can carry anti-dilution adjustments. Some warrants reprice or expand if a subsequent round clears below a threshold.
What it costs: even companies that recorded the warrant often recorded the wrong version of it. They entered the share number from the day of issue and never revisited it, so the record shows a fixed grant while the instrument itself grew with the next round. The cap table understates the dilution, the holder knows it, and the gap surfaces when the holder exercises for more shares than the record anticipated or when the buyer's counsel reads the adjustment clause and recalculates. The record has to reflect not just the warrant but its live terms, because a warrant whose coverage moves is a moving obligation, not a static one.
A warrant nobody authorized
A warrant is an agreement to issue shares in the future, which means it commits the company's equity exactly the way a share issuance or an option grant does. It therefore needs board authorization, and where the coverage is large enough or the terms unusual enough, sometimes shareholder authorization as well. Because the warrant rides inside a loan or a lease that the board approved as a financing decision, the equity-specific approval often never happens as its own act, and the minute book has no resolution that authorizes the warrant on its own terms.
What it costs: this is the same failure we described in who authorized this issuance, applied to an instrument that is easier to overlook. The buyer's counsel cannot find the approval, asks the seller to produce a ratifying resolution after the fact, and notes the gap in the disclosure schedule. If the warrant has already been exercised into shares, the question becomes whether those shares were validly issued, which is a heavier finding than a pending warrant because it touches the register and the certificate, not just a future right.
The expiry question nobody can answer
Every warrant has a term and an exercise mechanic, and both are diligence questions the company should be able to answer instantly and frequently cannot. When was it issued. When does it expire. Is exercise cash only, or is there a cashless or net-exercise option that changes how many shares actually come out. Are there conditions, such as a change of control, that accelerate or terminate it.
What it costs: the buyer builds a pro-forma fully-diluted table for the transaction and needs to know, for every warrant, whether it will be exercised into the deal, expire before it, or convert on the change of control itself. If the company cannot state the term and the mechanic from its own records, the buyer reconstructs them from the underlying agreements, and any warrant whose status is genuinely ambiguous, expired or not, exercised or not, becomes a point to negotiate rather than a fact to record. A warrant you cannot describe precisely is a warrant the other side gets to characterize for you.
Exercise, when a warrant becomes a share
The moment a warrant is exercised is the moment it crosses from a promise into the register, and it is the transition most likely to be recorded on only one side. The holder pays the exercise price, or nets it against the share value, and is entitled to a certificate and a register entry. What we see instead is the warrant marked exercised in a spreadsheet while the register is never updated, or shares appearing on the register with no cancellation of the warrant that produced them, so the same economic interest is now counted twice: once as an outstanding warrant, once as issued shares.
This is the warrant version of a problem we have written about for transfers, in how share transfers quietly corrupt ownership records. The fix is the same in principle. Exercise is a single event that must close the warrant and open the share issuance in one connected action: the warrant is marked exercised and removed from the fully-diluted count, the shares are issued with a board authorization and a certificate, and the register reflects the new holder as of the exercise date. One event, every artifact, one date.
What a clean warrant record looks like
A warrant does not need special handling so much as it needs to be treated as a first-class instrument rather than an attachment to someone else's contract. The record a company should be able to produce on demand, for every warrant it has ever issued, is short:
- The holder and the instrument. Who holds the warrant, how many shares it covers, the class, the exercise or strike price, and the transaction it was issued alongside.
- The live terms, not just the day-one terms. If the coverage is a percentage, an anti-dilution adjustment, or otherwise variable, the record carries the formula and the current resulting share number, updated after each event that moves it.
- The authorizing resolution. The board resolution, and any shareholder approval, that authorized the warrant on its own terms, filed in the minute book and referenced from the warrant record.
- The term and the exercise mechanic. Issue date, expiry date, whether exercise is cash or cashless, and any acceleration or termination triggers, so the warrant's status is a fact in the record rather than a reading of the contract.
- Its effect on the fully-diluted table. The warrant is counted in the fully-diluted cap table while it is outstanding, and on exercise it converts to a share issuance with a certificate and a register entry, and is removed from the outstanding count in the same action.
The theme running through all of it is the one we keep coming back to: the cap table is a view of the underlying records, not a record of its own, which we argued in your cap table is not your corporate record. A warrant proves the point more sharply than any other instrument, because it is precisely the claim that a spreadsheet cap table is most likely to omit and a proper record is built to carry.
The bottom line
Warrants are forgotten not because founders are careless but because the instrument is designed to live in someone else's paperwork. It is issued by a counterparty, drafted by their counsel, folded into a loan or a lease, and handed to a company that has no natural home for it between the share register and the option plan. The result is the one line the fully-diluted table is most likely to be missing, and the one a diligence reviewer is trained to go looking for in the debt file precisely because they know it is often absent from the cap table.
The prevention is not complicated. Record the warrant the day it is signed, with its holder, its coverage terms, its authorizing resolution, and its term, and count it in the fully-diluted picture from that day forward. If you keep options, SAFEs, and warrants in the same place as the shares they will one day become, there is no separate file to forget. Our cap table and equity and digital corporate records solutions are built so that a warrant is a tracked instrument with its own record and its own authorization, not an email you hope to find before the buyer's counsel does. The governance maturity assessment will tell you, in a few minutes, whether the warrants you have issued are actually on your fully-diluted table today.
