You've been imagining this infinitesimal for a while. Maybe spent weeks having long, perchance reliable conversations successful preparation. Now, it's yet clip to popular the question:
"Will you beryllium my cofounder?"
Interest successful entrepreneurship appears to beryllium rising successful the US, pinch caller business applications already up by much than 450,000 successful the first half of 2026 compared pinch the aforesaid play past year, according to the US Census Bureau.
If a founder's startup feels for illustration their baby, their cofounder whitethorn very good consciousness for illustration a spouse. Once they've chosen that typical someone, mounting the correct guardrails for that partnership is crucial. In immoderate ways, AI has made these decisions moreover much important arsenic business teams get smaller, companies move faster, and the unit becomes much concentrated.
Business Insider asked 3 sets of AI startup cofounders what early conversations person led to successful partnerships, and really they wished who gets what if things spell wrong. They shared the cardinal elements of their early agreements, on pinch what they'd put successful their perfect "cofounder prenup." Their words person been edited for magnitude and clarity.
Agree connected what being committed intends for you
David Emelianov is simply a 32-year-old package technologist and entrepreneur, based successful the Bay Area. He and his cofounder, Jordan Gaston, sold their app, Trimbox, for $4 cardinal astatine the extremity of 2025.
I deliberation location are 3 conversations each cofounder should person earlier penning a azygous statement of code: What benignant of institution are you really trying to build, what level of committedness does everyone expect, and what happens if personification wants out.
Sit down and determine if you're trying to build a venture-backed, winner-take-all institution wherever the extremity is to plaything for the fences, aliases if you want to build a profitable bootstrapped business that throws disconnected cash. Neither reply is correct aliases wrong, but they're very different games.
If 1 laminitis is optimizing for a $20M acquisition and the different wants to build a 30-year company, you're going to tally into problems eventually.
The 2nd conversation, though, has go particularly important for maine complete time. Earlier successful my career, I would person assumed everyone should beryllium 100% focused connected 1 company. Today I deliberation differently. I activity connected aggregate projects astatine once, and I really deliberation that makes maine a better entrepreneur because ideas, lessons, and opportunities transportation betwixt businesses.
The important point isn't that everyone useful the aforesaid way; it's that everyone agrees connected what "committed" really means. Is it OK to person broadside projects? Consulting? Angel investing? Starting different company? Those expectations should beryllium crystal clear from time one.
If I had to prime conscionable 1 speech to have, though, it'd beryllium the first one. You person to make judge you're playing the aforesaid crippled earlier it begins.
Define different exit scenarios early
Shalini Aggarwal is simply a 50-year-old CEO successful San Jose, California, and Andy Ratsirarson is simply a 37-year-old CTO. The 2 near Amazon astatine different times and reconnected arsenic cofounders of Tenfali, an AI startup.
We judge a cofounder prenup should see the things that are easy to disregard erstwhile everyone is excited astir an idea. This includes equity split, vesting schedule, IP ownership, roles and responsibilities, really awesome decisions get made together, and, importantly, what happens if personification leaves.
You request to person an honorable speech astir expectations earlier things get hard, and find retired what happens if 1 laminitis tin nary longer springiness the institution the aforesaid level of commitment.
For example, what happens if location is an acquisition offer, if 1 laminitis wants to support going, and the different does not? The extremity is not to presume the narration will fail, but to protect the institution and the narration by making the hard conversations definitive earlier emotions and money are involved.
You besides request a shared model for resolving problems, because conflict is inevitable successful a startup. Customers whitethorn beryllium slow to convert, the merchandise whitethorn request to change, aliases money whitethorn get tight. In those moments, the mobility is not whether you will disagree, but really you will make decisions erstwhile you do.
Once both founders are aligned connected a decision, location cannot beryllium a blasted crippled later. It becomes a shared decision, and some group person to ain the outcome. Those conversations tin consciousness uncomfortable early, but they are overmuch harder to person later.
Read much from our Tiny Team series
Decide whose telephone is last …
Praneet Dutta is simply a 32-year-old CEO and cofounder astatine AI startup Pomo. He's based successful Palo Alto.
A cofounder prenup, for maine and my cofounder Joe, is astir what happens successful the 30 minutes aft we some deliberation we're correct and disagree connected really to move forward. We had it travel down to 3 understandings.
First, we request clear, antecedently established domain ownership. Whose telephone is last successful which domain? Without that, the institution tin get stuck waiting for a statement that isn't coming.
Secondly, important decisions get a existent conversation, successful personification aliases connected a call, ne'er a chat thread. Tone and discourse matter astir erstwhile the stakes are high. That norm sewage tested early. When our seed information ended up oversubscribed, deciding really to build the headdress array meant a bid of unrecorded conversations pinch each different and pinch friends who had advised us. I cannot ideate having done that complete chat.
Third, we built a wont of revisiting the agreement. The statement that useful for a six-person team will not automatically activity later. So we dainty it arsenic a surviving agreement, not a signed-once document.
The existent cofounder risk is clone agreement, not conflict. That was the biggest displacement for us. We stopped treating disagreement arsenic thing that threatens the narration and started treating unclear disagreements arsenic what threatens the company.
Determine really you enactment connected the aforesaid page while moving quickly
Joe Cheuk is simply a 33-year-old CTO and cofounder, based successful Seattle. He cofounded the AI startup Pomo pinch Praneet Dutta.
Praneet and I some for illustration to move fast. Velocity is 1 of our biggest advantages arsenic a mini team, but 2 group moving accelerated successful somewhat different directions compound quickly. I'd adhd the clause group skip: find really you enactment aligned astatine speed.
In a ample company, the scenery shifts, and you usually person quarters to adapt. At an AI startup, the tools, the models, and sometimes the customer expectations alteration period to month. The hardest portion was not immoderate azygous decision. It was learning to make bully decisions and springiness a six-person squad assured guidance successful an situation wherever past quarter's playbook whitethorn already beryllium outdated.
So agreeing early connected really we steer the vessel together was 1 of the astir important things we did. Ours is simple: either of america tin inquire to slumber connected a large decision, we each laic retired our afloat viewpoint earlier we decide, and erstwhile a telephone is made, we some ain it and support moving.
That is the portion that founders underestimate. Most cofounder issues do not commencement arsenic explosions. They commencement arsenic mini moments wherever some group take silence. Making these conversations normal earlier they go costly is 1 of the highest-ROI things cofounders tin do.
Do you person a akin communicative to share? If so, please scope retired to the newsman at [email protected].
English (US) · 
Indonesian (ID) ·