You promised equity you can't afford to keep

You said "two percent, once we raise" over cold pizza in a borrowed office. You meant it. Now the term sheet is signed, the cap table is a Jenga tower, and two percent is a number you'd have to carve out of your own next round. The person you said it to still remembers the exact night. You mostly remember being terrified.

The words were spoken by a founder who no longer exists — the desperate one. The bill lands on the founder who's still here.

The dilemma

This is a composite of the kind we see in this founder category — assembled from many, not one real person.

Rohan bootstrapped for fourteen months. In month three, he pulled in Priya — a designer who took no salary, shipped the first product, and stayed when a paying gig would have been the sane choice. One night, exhausted and grateful, he said it plainly: "When we raise, you're getting two percent. You earned it." No email. No side letter. A handshake and a promise both of them believed.

The seed round just closed. The lead pushed the option pool wider, the founders' stakes compressed, and a fresh key hire wants a real grant. Rohan's own equity now has to stretch across a Series A he can already see coming. Two percent to Priya, on paper, is a chunk of the runway he's counting in shares.

Three forks sit in front of him. Honor the words — draft the 2% grant, take the dilution, keep the promise a scared version of him made. Quietly renegotiate down to 0.75%, reframing it as "that was aspirational, the market's different now, that's not how it was really framed." Or split the difference — offer 1.25%, call it fair, and hope she reads it as generous rather than as a promise getting shaved.

The ugly detail he won't say out loud: the renegotiation script is already written in his head, and it's persuasive. He knows he could sell it. He's a founder. Selling a smaller number is a Tuesday.

The read

Four of the eight lenses bite here. We'll walk them, then commit.

Honesty — is your account of the past true?

Honesty asks whether the story you're telling matches what happened. "That's not how it was framed" is the load-bearing lie in option two. It was framed exactly that way. Priya can quote the sentence. Rebranding a clear promise as a vague aspiration isn't a negotiation — it's editing the record after the fact and betting the other person doesn't have the original.

Honor the words: +4. The split lands at 0 — honest only if you say "I promised more than I can give and I'm asking you to absorb the gap," and dishonest the moment you dress it as fair-market. Quiet renegotiation: -4, because it requires rewriting a shared memory.

Duty — what does your role actually owe?

Duty asks what the position obligates, not what feels comfortable tonight. Rohan's role isn't "the guy who said a nice thing." It's the founder who accepted eighteen months of below-market work against a specific promise. The work was delivered. The obligation didn't evaporate because the cap table got tight — a tight cap table is a founder problem, not a Priya problem.

Here's the tension with honesty, and it's real: duty says pay what you owe, and the honest reduction ("I over-promised, here's what I can do") is genuinely truthful but still a partial default on the debt. Honesty can bless the split. Duty won't fully. Honor the words: +4. Split: -1. Renegotiate: -4.

Motive — what's really driving the appealing option?

Motive asks what engine is under the move you're leaning toward. Rohan admits the renegotiation script is already written and persuasive. That's the tell. The pull toward 0.75% isn't discernment about the market — it's the comfort of being able to win the conversation. When the reason you like an option is that you're good at defending it, the motive is self-protection wearing the costume of prudence.

Renegotiate: -4. Split, if chosen to look generous while paying less: -2. Honor the words: +3 — the motive there is simply "the debt is real."

Non-harm — who eats the irreversible cost?

Non-harm asks who carries the damage that can't be undone. Priya made an irreversible bet — eighteen months she can't get back, priced against your word. Shaving her grant transfers your dilution problem onto the one person who had the least protection and the most trust. The company survives either way. Her belief that your word means something does not survive a quiet haircut.

Renegotiate: -4. Honor: +3. Split: -1 — a reduction delivered honestly still costs her, just without the added injury of being lied to about it.

Verdict: Honor the 2% — and if the cap table genuinely can't bear it, go to Priya with the true sentence ("I promised 2%, I can fund 1.25% cleanly today plus the rest vesting at the next round") and let her decide with full information — but never renegotiate under cover of a rewritten memory.

A general AI chatbot will hand you whichever ending you tilt toward: prompt it while feeling broke and it validates the haircut; prompt it while feeling guilty and it validates the payout. Bending toward the user's lean is a known tendency of these assistants — they're tuned to keep you nodding in the moment. KarmaLens runs eight fixed lenses whose scores aggregate the same way no matter how you phrase the ask, and the verdict doesn't renegotiate itself to make you feel generous about paying less. When a verse is attached, it's the corpus text verbatim, cited — not a paraphrase bent to your mood. You can dislike the call. You can't talk the honesty score up to zero by explaining that the market's different now.

The takeaway — the two-founders test

You don't need us for this one tonight. The trap in a broken verbal promise is that a past you made it and a present you is stuck with it, so you quietly treat them as two different people with two different debts.

So separate them on purpose. Write two sentences.

• The founder who made the promise: "On [that night], I told [name] exactly this: ___." Use their words, not your later summary.
• The founder paying the bill: "Today I can fund this much cleanly: ___, and this much only by hurting: ___."

Now read the gap out loud. If your leaning move requires the second founder to rewrite the first one's sentence — to claim it was "aspirational," "not really a promise," "different framing" — stop. That's not negotiation; that's forgery of your own past. If instead your move keeps the first sentence intact and honestly says "I over-promised, here's the real number and why," you're allowed to renegotiate — because you're negotiating the payment, not the promise.

The line is simple: you can be honest that you can't afford what you said. You cannot be honest by pretending you never said it. Score just the leaning option on honesty and non-harm, and notice which one makes you flinch when you imagine reading it back to Priya's face.

Ready to run your own fork through all eight lenses? Take the founder decision through KarmaLens — or read a few worked verdicts in the gallery first.

When the debt is uncomfortable but clear, which is louder in you right now — the promise, or the case you've already built for shrinking it?