YC's Requests for Startups Lists 16 Ideas and Not One Is Consumer. Should You Apply to YC F26 With a Consumer App?

YC's current Requests for Startups lists 16 ideas the partners want founders to tackle. Count the consumer ones.

There are none.

The Summer 2026 list opens with "AI has stopped being a feature and started being the foundation," then runs through inference chips for agent workflows, counter-swarm drone defense, semiconductor supply chains, AI-native service companies, company brains, personalized medicine, and industrial capabilities on the moon. Sixteen entries. Zero aimed at ordinary human beings using a product they like.

The two that come closest, AI Personalized Medicine and Dynamic Software Interfaces, both end up framed as infrastructure and delivery primitives rather than products a person chooses to open. Even the near-misses are pitched at builders, not users.

With the F26 deadline on July 27 at 8pm PT, that is worth staring at if you are building for consumers.

Does YC fund consumer startups in 2026?

Yes. But read the room.

Start with what the RFS actually is, because YC tells you at the top of the page. These ideas "represent just a fraction of what we fund," and "you don't need to work on these ideas to apply to YC."

That is not boilerplate, it is the literal answer. Consumer is not banned. No filter drops your application because you did not pick an item off the list. YC funds off-RFS companies in every single batch.

But the RFS is still the best public map of where partner attention currently sits, and consumer is not on the map. You are applying against the grain. Price that in rather than pretend it away.

The line consumer founders should sit with

Inside Aaron Epstein's RFS entry, "Software for Agents," is this:

"The next trillion users on the internet won't be people, they'll be AI agents. And now is the time to 'Make Something Agents Want.'"

That is a deliberate riff on the sentence printed in the footer of every page on ycombinator.com: Make something people want.

YC's institutional thesis for this cycle is that the growth market in users is no longer human. A consumer application to F26 is, structurally, a bet that the motto still outranks the memo.

It is a defensible bet. Just know you are making it.

The counter-evidence ran on YC's own forum this week

On Thursday a developer posted a daily anagram puzzle called 18 Words to Show HN. No AI. No enterprise buyer. No agent harness. A word game.

As of today it has 1,125 points and 355 comments, the tenth highest-scoring Hacker News post of the month so far. In the thread, the developer casually posted their own analytics: "only 32 people out of 12,145 got all 18 in todays puzzle so far." Twelve thousand people played a single day's puzzle.

Less than three hours after that word game went up, a YC S26 company posted its Launch HN on the same page. Context.dev, an API that turns any website into structured data for AI agents. Exactly the kind of company the RFS is asking for.

It sits at 117 points.

Same forum, same morning: 1,125 Hacker News readers upvoted the word game, nearly ten times as many as upvoted a funded AI startup built squarely on the RFS thesis.

That is not a prediction about who builds the bigger business. Context.dev may well win by a mile. It is a fact about attention, and a useful corrective to the idea that nobody wants consumer products anymore.

YC's biggest outcomes were consumer

YC's apply page names the founders who often come back to tell the batch their origin story: OpenAI, Airbnb, Stripe, and DoorDash. Two of those four are consumer marketplaces.

Reddit was in YC's first batch ever (S05). Airbnb (W09) was three guys renting air mattresses. DoorDash (S13) was a restaurant delivery page with a phone number on it. Instacart (S12) and Coinbase (S12) both went out consumer-first.

And this is not ancient history filed away in a museum. Two of the partners with bylines on that very RFS list built consumer companies themselves: Gustaf Alströmer ran growth at Airbnb, and Tom Blomfield founded Monzo. The people who wrote the AI-infrastructure wish list are the same people who will instantly recognize a working consumer product when the numbers are in front of them.

What a consumer application must prove that a B2B one doesn't

Here is where founders throw the application away. They read "YC wants AI infra," and respond by costuming a consumer product as an AI platform. Partners have seen that outfit a thousand times, and it converts an honest consumer application into a B2B application with a lie in it.

Do the opposite. Lean all the way in, and win on the one axis where only consumer can win.

A B2B application is graded on the wedge: who is the buyer, what do they pay, why now, why you.

A consumer application is graded on retention. That is the whole thing.

Which is exactly why the 18 Words number is a warning as much as a flex. Twelve thousand players is a Hacker News spike, not a business. The only question a partner will ask about it is: how many came back the next day? A front page is distribution you were loaned. Retention is distribution you own.

Four things to put in the F26 form

• "How far along are you?" Retention by cohort. D1, D7, D30. If the curve flattens at 25% and stays there, write that number and stop talking. It does more work than any paragraph you could write.
• "Why now?" For consumer in 2026, the honest why-now is almost always distribution. Search is being eaten by AI answers, app stores are saturated, feeds are drowning in generated slop. If you have found a channel that still works organically, that is your why-now, and it is a sharper one than most AI startups have.
• "How will you make money?" This is consumer's weakest link and the exact place applications go vague. Do not go vague. Say the thing, even if the thing is "not yet, and here is the retention number we would need first."
• Never position yourself against the RFS. Don't write a defensive sentence about not being on the list. Just show the numbers.

Have someone push on the weakest line

The specific failure mode for consumer applicants is an application that reads charming and proves nothing. The words go to the idea, because the idea is the fun part, and the retention data ends up in a subordinate clause on line 40.

That is what YC Roaster is for. You get instant AI feedback trained on thousands of YC applications, and strong applications get matched with a YC alum who tells you which line a partner will stop on. It's free, and with two weeks to the deadline it is a cheaper way to find the hole than an interview in September.

The bottom line

Sixteen requests, zero consumer, and one of them arguing that people are no longer the growth market in users. That tells you where the tailwind is, and it is not behind you.

It does not tell you the door is closed. It tells you that you get no free credit for your category, and that you will be graded on a number rather than a story.

Meanwhile, 12,145 people played a word game, and 1,125 Hacker News readers upvoted it, nearly ten times as many as upvoted a funded AI startup that launched the same morning. People still show up for things they like, and YC has never stopped funding that. The question was never whether people showed up. It is whether they came back the next day, and whether you can prove it in a form field by July 27.