Y Combinator
Make something people want.
Start new thread

Y Combinator offers 7 startups ideas they want to fund (Spring 2026)

by
•

As usual, came up with segments that are worth investing:

1. Cursor for Product Managers
2. AI-Native Hedge Funds
3. AI-Native Agencies
4. Stablecoin Financial Services
5. AI for Government
6. Modern Metal Mills
7. AI Guidance for Physical Work 
8. Large Spatial Models 
9. Infra for Government Fraud Hunters 
10. Make LLMs Easy to Train

Which of the following do you think are the most interesting?

OR

Would you add to the list, maybe something that is more worth funding?

Also, are you going to apply? 👀

5.3K views

Add a comment

Replies

Best

What about an AI-native investment bank or fundraising broker?

A few players such as Metal, DeepFlows, Signal are heading towards this positioning helping startups raise money against a subscription at first, and eventually taking a success fee out of the deal. What do you think?

 It sounds fancy. How it should work according to you?

 the platform handles matching, warm intros, outreach, tracking. Only a small, curated subset gets human advisory on top.

Model:

-Platform users = subscription

-Assisted startups = success fee (outcome based)


This model helps both ideation/pre-seed startups in discovery, matching with incubators/early investors, as well as VC-backed or incubator-backed startups with their next rounds

 This idea doesn't sound bad. The thing is, who will make it happen? Are you about to create something like that? :)

Agentic AI safety and security are what will make all the sectors in the list investable, and we are really behind at the moment, so for me that's the number one sector worth funding.

 I agree with this one. Just let's have a look at how soft something is without knowledge, when it is vibecoded. It will break in a millisecond. Data leaks guaranteed. Are you building something in this sphere?

 Thanks Nika. Yes, I launched yesterday, but with problems and it got buried:
Not a big deal...PH is not best place for open source dev tools anyway :)

 Sorry to hear that. Maybe relaunching on other platforms could help with link building and partial visibility. :)

#1 really resonates — Cursor for Product Managers is exactly where we’re seeing the biggest gap right now.

At we’re building an AI-native product development loop that starts with guided intent capture (persona interviews → clear stories, scope, estimates, and outcomes) before anything gets built — not just faster docs, but better decisions.

AI is crushing execution speed, but what teams struggle with now is what to build, why, and how to continuously adapt once real usage shows up. PMs need AI copilots for thinking, not just writing tickets.

Feels like this category is going to define the next generation of product tools.

 Have you launched it yet? Or are you about to polish and launch? :)

 We launched on Product Hunt earlier this week. The INVENT product has been available since early January.

Hi! The list is very interesting, I have applied for Spring of 2026 and I think my startup fits in the AI Guidance for Physical Work category. I'm working on an AI tool to intended to help you take steps to improve tree health, and keep track of tree health, so as to achieve better outcomes, yield improvement (in case of fruit trees), and climate resilience (in some cases), better tolerance to heat and stress (mitigate climate change). Essentially like Google Analytics for trees so we have that continuous feedback loop from the tree in the form of images, and soil data so repeated interventions can help longer term better outcomes.

 The question is: How do you want to monetise this business?

We have to charge a license fee per tree - for orchards. The incremental yield and reduced fertilizer cost will justify it.

Isn't " for Product Manager" already available - as Cowork?

   Isn't Claude cowork for more general tasks? At least I perceive it that way.

   It was, until they launched the on 30th Jan, which essentially are those field's specific expertise tools that caused software native companies' shares to tank, all around the globe!

   In that case, some closer clarification by YC could be cool because at this point I do now know what "innovative" they want to see in these terms.

They seem to have updated the article and I really like Infra for Government Fraud Hunters - even though mainly focused on the US, pretty sure that in EU, especially central / eastern countries the fraud and corruption is quite high.

Also really like the push to make the LLMs easier to train!

 Thank you for the update, yeah, they added a few more items, gonna add them to the list :)

#1 and #3 hit different for me.

Cursor for PMs is wild because most PM tools still feel like 2015 — imagine actually managing sprints, standups, and priorities through AI right in your IDE. That's where we're headed with GitScrum MCP Server (literally why we built it).

AI-Native Agencies is the sleeper hit. Agencies are drowning in client work, and autonomous agents could flip the whole model — less burnout, better margins, faster delivery.

What I'd add: AI-native dev tools that talk to your entire stack — not just code, but PM, CRM, time tracking, all through one protocol. MCP is early proof of concept.

Not applying yet, but watching closely 👀 What about you? Building something in one of these spaces?

 Interesting POV. Maybe they will add items to the list (at least, they extended it from 7 to 10 items). And no, I do not have ambition to apply :D

The #1 on this list really resonates with me "Cursor for Product Managers."

I'm actually building in this exact space with . The idea: you paste a URL and it automatically crawls the site, captures every screen, detects the navigation flows, and generates an interactive visual map you can annotate and analyze with AI.

Think of it as what Cursor did for developers, but for designers and PMs doing UX benchmarking. Right now, most product teams are still screenshotting competitor apps and pasting them into Miro or Figma. It's manual, slow, and impossible to keep up to date.

 Will you send an application to Ycombinator?

 I’ve thought about it, but I’m probably not applying this batch. It’s early, I’m solo right now, and YC tends to prefer teams. I’d rather get some traction first and revisit it. Open to "co-founders" though 🤣

 aaa, okay, you are valid in both points :D

 Would you use a tool like as a marketer? 🤓 I can see clear use cases for PMss and Product designers but I'm curious whether marketers do this kind of competitive analysis of research on other webs or products.

#1 is interesting. But it needs to address how the PM role itself needs to change with AI. Scrum-based planning misses the new super-power - multi-phase planning. A chat-based tool that can create PRDs (and keep them alive by managing change requests the same way we manage Git) in a structured data format JSON) that eg Cursor but platform-agnostic can read via MCP to formulate its own plan (asking clarifying questions, approval gates for plans etc) would be the game changer. AI Jira isn't exciting (Ask Rovo). PMs need to move upstream from scrum micro-planning to focus on the requirements layer - Requirements As A Service will be an entirely new service category. That would be interesting.

 If AI can work with complexities (yes, it can), it would improve the whole process, of course... the thing is whether teams can work at such a pace. I still somehow feel the human nature to be prepared for huge changes.

Oh no wonder i got rejected from YCombinator😩 ...... My product isn't in this list.

My Vote for the list would be [2], [5], [4], [6]

I would add AI to create assessment from Learning materials & video

I dont understand the other product category

 What was your product you tried to apply with?

First
Previous
123
Next