---
title: "YC's \"Own Your Intelligence\" Request for Startups, Explained"
date: "2026-10-30"
video_id: "7702553360460729613"
tiktok_url: "https://www.tiktok.com/@y_combinator/video/7702553360460729613"
creator: "@y_combinator"
category: "AI"
tags: ["y combinator", "request for startups", "open weight models", "fine tuning", "ai startups"]
excerpt: "YC's Own Your Intelligence request asks for startups that help companies own specialized models. What it says, what it backs up, and what it leaves unsourced."
---
"Own Your Intelligence" is an entry in Y Combinator's Winter 2027 Requests for Startups, written by Jon Xu, asking for startups that help companies train and own specialized AI models instead of renting frontier models. A TikTok from @y_combinator reads the request almost word for word. Its central claim, that open-weight models are "competitive with the top frontier models on meaningful benchmarks", names no benchmark or model, and the cost and history claims around it carry no figures.

## What is YC's "Own Your Intelligence" request?

The [YC Requests for Startups page](https://www.ycombinator.com/rfs) lists "Own Your Intelligence" under the Winter 2027 heading, with Jon Xu as author. The page shows no publication date. A [FourWeekMBA write-up](https://fourweekmba.com/ai-yc-names-11-startup-ideas-as-ai-bottlenecks-move-past-models/) dated 2026-10-28 quotes its opening line, "Today, nearly every company building with AI rents its intelligence." The video, posted 2026-10-30, opens with the same sentence.

The video's lower third begins "Jon X" and "General", which fits a title card for Xu. The YC page calls him a partner.

## What is the argument?

The monologue runs in four steps:

1. Companies rent intelligence from frontier labs.
2. That "was an easy trade when the best models were years ahead of everything else," and "nothing you did with your own data could close the gap."
3. Open-weight models have now caught up on "meaningful benchmarks."
4. Companies will want to own specialized models trained on their own workflows, data and failures.

The video does not argue that renting ends. It says "Companies will keep renting frontier intelligence for their hardest problems."

## Are open-weight models really competitive with frontier models?

The video says "Open weight models are now competitive with the top frontier models on meaningful benchmarks." It names no model, benchmark, score or date, and neither does the YC page text or the FourWeekMBA summary.

The wording is hedged. "Competitive" on "meaningful benchmarks" does not claim open models match frontier models on every task, and the later line about renting for the hardest problems is consistent with that. To test the claim yourself, pick the benchmarks that match your work and compare current open-weight and frontier results from a leaderboard or model card.

## When were frontier models "years ahead"?

The video gives no year, model or size for that gap. The claim explains why the economics have supposedly changed, but nothing in the video or the sources dates when the gap closed.

## Is renting AI more expensive in regulated industries?

The video says the issue "matters most in regulated industries, where strict data governance and lockdown infrastructure make renting even more expensive for the performance they get." It names no cost figure, industry, regulation or customer, and the YC page does not add numbers. Read it as YC's judgment.

## What kind of startup is YC asking for?

The most concrete passage lists the work: "Someone has to turn workflows into evals, expertise into training data, production failures into improvements, and carry all of it forward with each new base model release." That is four jobs:

- building evaluations from a company's real workflows
- turning expert knowledge into training data
- feeding production failures back into training
- moving all of it to each new base model when one is released

The video closes with "If you're helping companies own their intelligence, we want to hear from you." It gives no URL, deadline or example company, and it shows no demo of a model or training run. The YC page is where the request lives.

## Is "own your intelligence" a YC-only idea?

No. Search results surfaced related commentary from Sequoia, also titled "Own Your Intelligence", and from Fireworks AI arguing a similar thesis. I did not read those pages, so they show only that other firms use the phrase.

## Bottom line

The video is YC reading its own Winter 2027 request: help companies turn workflows, expertise and failures into models they own. The four jobs it lists are a concrete description of the opportunity. The supporting claims, that open-weight models are competitive on benchmarks, that frontier models were once years ahead and that regulated industries pay more to rent, come with no benchmarks, dates or figures.

## Sources

- [Requests for Startups, Y Combinator](https://www.ycombinator.com/rfs): the "Own Your Intelligence" entry under Winter 2027, by Jon Xu.
- [YC Names 11 Startup Ideas as AI Bottlenecks Move Past Models, FourWeekMBA, 2026-10-28](https://fourweekmba.com/ai-yc-names-11-startup-ideas-as-ai-bottlenecks-move-past-models/): secondary report with the opening quote, author and batch.
- [TikTok video by @y_combinator](https://www.tiktok.com/@y_combinator/video/7702553360460729613): the video reviewed here.


## Video description and verification notes

**The file is an 88.755 s video (audio 88.723 s), 1080x1920 (9:16), MP4 (isom) container, HEVC Main video at 2997/125 fps (about 23.98) and 519,276 bps, plus HE-AACv2 stereo audio at 44.1 kHz and 64,117 bps. The container bitrate is 588,378 bps and the file is 6,527,722 bytes. It was posted at epoch 1793390479, which is 2026-10-30 in UTC (the upload_date field says 20261030), by @y_combinator (nickname "Y Combinator", uploader id 7019683670089237510). The audio is credited as "original sound" by Y Combinator. The description is 296 characters and opens with the same words as the title, which TikTok truncated. Counts at capture (2026-10-31T12:08:42Z) were 47,300 views, 2,252 likes, 45 comments, 306 reposts and 804 saves. I read one frame, frame_001 (0 s), and no others, so I cannot describe the rest of the visuals. Frame_001 shows a man at a wooden table in a blue collared shirt against a red patterned wall, hands clasped, with a burned-in caption "Today, nearly" and a white lower-third that was still animating in: "Jon X", "General", "Partner,". The fragment is consistent with a title card for a General Partner named Jon X..., but the surname never fully rendered in the frame. The transcript has 18 segments and about 270 words. It is a single monologue arguing that companies rent frontier-lab intelligence, that open-weight models have closed the gap, and that companies will want to own specialized models trained on their own workflows, data and failures. The last segment runs from 83.32 s to 88.40 s, which is inside the 88.72 s audio, so no segment was fabricated past the end of the audio.**

## The video is a read-out of a Y Combinator Request for Startups, written by Jon Xu

The spoken text matches a Winter 2027 Request for Startups entry titled "Own Your Intelligence." I fetched the [YC Requests for Startups page](https://www.ycombinator.com/rfs), which lists the entry under the "Winter 2027" heading with Jon Xu as author. The page shows no publication date. A FourWeekMBA write-up, which I used only as a secondary report, quotes the opening line "Today, nearly every company building with AI rents its intelligence." The transcript begins with the same sentence at 0.00 s. That secondary report is dated 2026-10-28 and says the list appeared on the Wayback Machine on 2026-10-27, so the video (posted 2026-10-30) came after the written request. I could not independently confirm the Wayback date. The video therefore adapts existing text and did not originate the claims.

The frame supports the identification only partly. The lower third shows "Jon X" and the title word "General," and the page calls Xu a partner. I did not read a frame where the full name or title finished rendering, so I rely on the YC page for the full name and treat the "General Partner" reading as consistent but unconfirmed. The YC page as fetched named him only as a partner, and I did not verify the exact title.

## "Open weight models are now competitive with the top frontier models on meaningful benchmarks" is an assertion with no benchmark named

At 19.40 s the video says "Open weight models are now competitive with the top frontier models on meaningful benchmarks." The only frame I read shows nothing of a chart, a model name or a leaderboard, and the transcript names no model, benchmark, score or date. I did not find a primary benchmark source in this research, so the claim is Unverified here. The FourWeekMBA summary and the YC page text state the same claim in the same generality. A reader who wants to test it would need to pick specific benchmarks and compare current open-weight and frontier results from a leaderboard or model card. I did not do that, and I make no claim about which models lead.

The sentence also carries a hedge worth noting. It says "competitive" on "meaningful benchmarks" and does not say open models match frontier models on every task. The later line at 74.16 s, "Companies will keep renting frontier intelligence for their hardest problems," concedes that frontier models remain the choice for the hardest work. The two statements are consistent with each other.

## "Renting makes sense when frontier models were years ahead" is a historical claim with no date

At 10.48 s the video says the rental trade "was an easy trade when the best models were years ahead of everything else." It gives no year, model or gap size. I found no source in this run that dates when open-weight models closed the gap, so the claim is Unverified. It is framed as the reason the economics have changed, and it also underlies the argument that "nothing you did with your own data could close the gap" (14.68 s). Neither point is measured in the video.

## The regulated-industry cost claim is an opinion with no figures

At 43.56 s the video says the issue "matters most in regulated industries, where strict data governance and lockdown infrastructure make renting even more expensive for the performance they get." No cost figure, industry, regulation or customer appears in the transcript or in the frame I read. The YC page summary I fetched also names regulated industries without numbers. I treat this as a stated judgment by YC and mark the cost comparison Unverified.

## The startup opportunity is described concretely, but no company, product or demo appears

The most specific passage is at 58.16 s to 69.20 s: "Someone has to turn workflows into evals, expertise into training data, production failures into improvements, and carry all of it forward with each new base model release." That lists four jobs: building evals from workflows, building training data from expertise, feeding production failures back into training, and migrating the work to each new base model. The description repeats the same idea. The closing call at 83.32 s to 88.40 s says "If you're helping companies own their intelligence, we want to hear from you." The video gives no URL, no application deadline and no named startup, and the only frame I read has no on-screen link. The YC page is the place the request lives, and it names Winter 2027 but does not give a date in the version I fetched. The video also shows no demo of any model, training run or tool in the frame I read.

The search results also surfaced related commentary from Sequoia ("Own Your Intelligence") and Fireworks AI that argues a similar thesis. I did not fetch those pages, so I do not cite them as evidence for any claim here. They indicate the phrase is used by other firms, which means the framing is not unique to YC.

## Key Takeaways

- **Verified:** The spoken script matches the "Own Your Intelligence" entry by Jon Xu in YC's Winter 2027 Requests for Startups, including the opening sentence at 0.00 s and the closing call for founders.
- **Clarification:** The video is a read-out of existing written text. The secondary report dates the list to late October 2026, before the 2026-10-30 post, but the YC page itself shows no date.
- **Unverified:** "Open weight models are now competitive with the top frontier models on meaningful benchmarks" (19.40 s) names no benchmark or model, and I found no primary benchmark source in this run.
- **Unverified:** The claims that frontier models were "years ahead" (10.48 s) and that regulated industries pay more for renting (43.56 s) carry no figures or dates, in the video or in the sources I read.
- **Context:** The only concrete deliverables named are four jobs: workflows into evals, expertise into training data, production failures into improvements, and carrying work across new base model releases. The only visual evidence I read is one frame, so what the rest of the video shows is unknown.

## Resources

- [Requests for Startups, Y Combinator](https://www.ycombinator.com/rfs) establishes that the "Own Your Intelligence" entry sits under Winter 2027 and is by Jon Xu, with no page date.
- [YC Names 11 Startup Ideas as AI Bottlenecks Move Past Models, FourWeekMBA, 2026-10-28](https://fourweekmba.com/ai-yc-names-11-startup-ideas-as-ai-bottlenecks-move-past-models/) is a secondary report that gives the opening quote, the author, the batch and the Wayback timing. It cannot establish the original date by itself.
- [TikTok video by @y_combinator](https://www.tiktok.com/@y_combinator/video/7702553360460729613) is the video analyzed. I did not fetch it, and I used only the staged artifact files.
