What the default path produces
The default is that YC announces you. Its account has 1,644,637 followers, which sounds like the best distribution a seed-stage company will ever be handed for free.
We track every AI launch we can find on X. Across 1981 launches measured, the 23 companies announced from that account have a median of 54,268 views. The range runs from 16,348 to 2,450,323, and the top of that range is one outlier carrying the average.
For comparison, a company with between 1,000 and 10,000 followers posting its own launch has a median of 318,959 views. That is 190 launches, so it is not a small sample. Announcing yourself with five thousand followers beats being announced to 1.6 million by roughly six times.
YC is doing something useful here, and it is free. The point is about how the platform works.
A roundup post from an institution and a founder's own post are different objects. One is a feed obligation, the other is a person with a thing they made.
The batch day is most of the problem
Look at when that account posts. When it announced one company on a given day, the median was 97,212 views. When it announced several on the same day, the median was 26,496.
Sharing your announcement day with two or three batchmates costs about three quarters of your reach. The mechanism is the same one that runs the rest of the platform.
Attention is winner-take-most inside a topic on a given day, and if the topic is "YC companies launching", you are competing with the people sitting next to you.
The same effect shows up at a larger scale. On 3 September, OpenAI announced GPT-6 Astra and took 88.6 million views. The thirteen other companies who launched that day had a median of 159,156, against 312,971 for the thirteen that launched the day before.
The YC company that beat it by 380 times
Tavus went through Y Combinator. On 1 October 2026 it launched Griffin, a video model that passes for a person, from its own account, which had 11,500 followers.
The post stands at 21.1M views, about 380 times the median the YC account produces for a batch company, with 5,634 quote posts and 30,000 organic posts across the timeline on launch day.
Doomers ran the amplification: 150 founders, creators and builders carried the post in the first hours, in an order planned before it went up. The case study has the hour-by-hour reads.
The same thing works at the small end. Guide Labs went through YC W24. When it launched Clarity, it posted from its own account, which had 1,504 followers, and the post did 570,904 views.
That is 1,093 times fewer followers than the YC account and 10.5 times the median reach that account produces for a batch company. The follower count did almost none of the work. Something else did.
What did the work is that other accounts carried it. That is the only mechanism on X that produces reach a company does not already own, and it is available to a company with 1,504 followers and unavailable to a company with a million if nobody chooses to pick the post up.
What I would actually do
If I were in a batch right now, this is the plan I would run.
Take the YC post, and treat it as one of several inputs. It costs nothing and it is a credibility marker that other people quote later. Just do not build the launch around it, because the median says it will do 54,268 views and your launch needs more than that.
Move off the batch day. Ask when the others in your group are going, and go on a different one. This is the single highest-return decision available and it costs a week at most.
Tuesday is the strongest day in our data at a median of 426,288 views across 146 launches, against 194,395 for Thursday across 113. Roughly 2.2 times, for changing nothing but the date.
Spend the twelve weeks in two buckets: story and distribution. Story is the video: the product doing the thing, shown in a form people repost. Distribution is the amplification: who carries the post, in what order, inside the first hour.
The story earns the attention and the amplification decides how far it travels, and that part is black and white. A strong film with no amplification stalls at the following. A strong film with amplification is Tavus.
The distribution bucket is relationships. The accounts that will carry your launch are people who already post about your category.
You have twelve weeks to become someone they recognise. Reply to them, be useful publicly, ship things they can see. A cold ask on launch day converts badly; a familiar name converts.
Write the post yourself, and make it specific. Guide Labs said "the first inherently interpretable AI platform is finally here" and named the product.
It is a claim with an edge on it. Most batch launch posts describe a category and could be swapped between four companies without anyone noticing.
Ship something people can open. The largest launch we have ever measured, GPT-6 Astra at 88.6 million views, had a below-median save rate because it went out as a limited preview and nobody could use it that day. If your launch points at a waitlist, the attention has nowhere to go.
Line up the first hour. Reach compounds from early velocity. Knowing in advance which ten accounts are going to engage in the first thirty minutes is worth more than any amount of polish on the asset.
The simplest tool for it is a calendar invite for the posting minute, sent to every investor, advisor, friend and teammate, 50 to 100 people, with the link added when the post is live. Everyone in your ecosystem should know you are launching, and when.
What I would skip
- Launching on Demo Day week. Every investor and journalist watching your category is looking at a hundred companies at once. It is the worst week of the year to need individual attention.
- The press embargo. A tier-two write-up published at 6am does less for a seed-stage AI company than four accounts your buyers already follow posting about it. Coverage follows attention now, in that order.
- Shipping the film without the distribution. A produced launch film is worth making. Make it, amplify it through the accounts your buyers follow, then put your own network behind it with the calendar invite. The film is the story. The amplification is what carries it.
- Posting and hoping. What separates a launch that travels from one that does not is almost entirely who was lined up beforehand.
Questions we get asked
Should a YC company let YC announce its launch?
Take it, and plan as though it will not carry the launch. Across the 23 companies announced from the YC account, the median is 54,268 views, against 318,959 for companies with 1,000 to 10,000 followers announcing themselves. It is a free credibility marker rather than a distribution plan.
When should a YC company launch?
On a day none of your batchmates are using, and preferably a Tuesday. Companies announced alone from the YC account did a median of 97,212 views against 26,496 for those sharing the day. Across all the launches we measure, Tuesday runs at a median of 426,288 and Thursday at 194,395.
How many followers does a startup need to launch well?
Fewer than most founders assume. Tavus launched Griffin from 11,500 followers and did 21.1M views.
Guide Labs launched from an account with 1,504 followers and did 570,904 views. Reach on X comes from who repeats a post rather than who receives it first, which is why follower count predicts launch performance so weakly.
How do you get engagement in the first hour of a launch?
Send a calendar invite for the posting time to every investor, advisor, friend and teammate, 50 to 100 people, and add the link when the post is live. That puts the whole network on the post inside the window X uses to decide distribution. The replies and saves from people who know the company give strangers a reason to talk about it on their own, which is where the reach past the following comes from.
Is Demo Day a good time to launch a product?
No. Demo Day concentrates every investor and reporter covering your category on a hundred companies in the same window, which is the opposite of the conditions a single launch needs. Ship the product announcement before or well after it.