What stealth costs you on the day
Stealth is usually the right call while you build. The bill arrives on reveal day, and it is larger than most teams expect.
New to the term? What a stealth startup is, and what the stealth launches in our directory did covers the definition and the numbers. This guide covers the launch.
Ranking systems judge a post relative to the account's own baseline. An account with no posting history has no baseline, which sounds neutral and is not: there is no evidence that this account's posts get responses, so the system has no reason to push the first one.
A company account posting its reveal cold is competing against its own standing start, and it gets capped inside the first stretch of life no matter how good the copy is. The mechanism is here.
The second cost is that nobody is waiting. A company that has been posting for two years reveals a product to an audience already interested in the problem. A stealth company reveals it to strangers who have to be persuaded, in one post, that the problem matters and that this is the answer.
The answer to nobody waiting is to make sure your own ecosystem is. Set up a calendar invite for the reveal and send it to every investor, advisor, friend and teammate, 50 to 100 people.
Everyone who knows the company knows the minute, opens the post when the link lands in the invite, and gives the account a first ten minutes it has never had. That early, real response is what strangers see when the feed starts testing the post on them.
The third is that there is nothing to search. Somebody who hears your name that day and looks you up finds a landing page and no context, at exactly the moment their interest is at its peak.
The advantage it gives you
Stealth buys one thing that no established company can buy back: nobody has already decided what you are.
A company that has been talking for two years has an existing frame. Readers know roughly what to expect, which means the announcement is judged against that expectation and mostly confirms it.
You have no such constraint. Whatever you say first becomes the category you are in, the comparison set you are judged against, and the sentence people use when they describe you to someone else.
That is worth spending real time on, because you get one attempt at it. Narrative design covers the work in full, and it matters more here than in any other launch type.
The related advantage is that time in stealth usually means real substance. Teams that have been building quietly for two years have benchmarks, artefacts and specifics that a company shipping every fortnight does not have accumulated. That material is the thing that travels, and most stealth reveals leave almost all of it in a folder.
Warm the founder, not the company
The fix for the cold-start problem is to stop the account being cold, and the account to warm is the founder's.
Personal accounts consistently outperform brand accounts on both reach and trust, and a founder can post about the problem space for weeks without revealing anything about the product. What they are building towards stays private. What they think about the domain does not have to be.
- Four to six weeks out, the founder starts posting on the problem, not the product. Opinions about the domain, observations from building in it, disagreements with how others frame it.
- Two weeks out, the posting is regular enough that the account has a real baseline and a small number of relevant people follow it.
- Reveal day, the founder's account carries the announcement and the company account quotes it. Both accounts benefit, and the one with history is the one carrying the load.
Two weeks is workable if that is what you have. What matters is that the account has posted enough for the system to have a sense of it, and that a handful of people who will amplify on the day already follow it before the day arrives. A forty-minute daily budget is enough to get there.
The worked example: TypeSafe AI
Doomers launched TypeSafe AI out of stealth on 15 September 2026. Two years in stealth, a founder account at 1,500 followers and a company account at 200.
Diogo Almeida co-invented RLHF and ChatGPT at OpenAI, which was the one credential a stranger could verify, so the post led with it, then the question the company exists to answer, then Jev.
The video ran nearly three minutes, and Doomers seeded it to 80 to 100 hand-picked engineers, founders, builders and creators, who carried the post in its first hour.
The post stands at 40M views, 16 of the 1978 launches in the directory, and the founder's account is at 154,182. Every step in this guide is on the TypeSafe AI case study, with the day-by-day numbers.
What to lead with
The most common stealth-reveal mistake is announcing existence. "After two years in stealth, we are excited to introduce" is a sentence about you, published to people who do not yet care about you.
Lead with the specific thing it does, demonstrated. The task it completes, the artefact it produces, the number that surprised your own team. Let emerging from stealth be a clause somewhere in the middle, where it belongs.
| Instead of | Lead with |
|---|---|
| Two years in stealth, now launching | The task it does end to end, named exactly |
| The team's pedigree | The artefact: a diff, a benchmark, an output somebody can inspect |
| The category you are entering | The specific segment you are best at, described precisely enough that a reader recognises themselves |
| The funding, if you have announced it | What you built with it. The round is a clause |
| A polished film | A screen recording, with the failures left in. Usually cheaper and better |
The week after is the launch
The most expensive stealth mistake is treating the reveal as an event rather than a starting point. A founder who goes from a handful of followers to a few thousand and then posts nothing for six weeks has converted a compounding asset into a screenshot.
Impressions peak in the first 48 hours and taper. The audience does not.
Those people now receive your next post natively, with no amplification and no budget, which is the actual prize. The view count is a proxy for it at best.
If you are revealing to a waitlist or early access rather than general availability, the follow-through matters more still, because the reveal produced a queue rather than users and the queue needs feeding.
One client added more than 10,000 waitlist signups in the four days after their launch, and what decided whether that was worth anything was what reached those people in the fortnight afterwards.
Plan the month after the reveal before you plan the reveal. It is the part almost nobody scopes, and it is where a stealth launch either compounds or evaporates.
Questions we get asked
Should we build an audience before coming out of stealth?
Yes, on the founder's account rather than the company's. Ranking is relative to an account's own baseline, so a profile with no posting history has no baseline to outperform and gets capped early no matter what it says. Four to six weeks of a founder posting about the problem space, without revealing the product, gives the reveal something to land on.
What should a stealth company lead with on reveal day?
The specific thing it does, demonstrated. Existence is not news to anyone outside your investors, and a post whose entire content is that a company now exists gives a reader nothing to react to. Lead with the task, the artefact and the number that surprised your own team, and let the fact of emerging be a clause rather than the headline.
How long before a stealth launch should we start posting?
Four to six weeks is comfortable and two is workable. What matters more than duration is that the account has posted enough times for the system to have a sense of how its posts perform, and that a few people who will amplify on the day already follow it before the day arrives.