The first thing to understand about Sky Yang is that he has been running the same play since he was a teenager. Find a group of people who share a problem. Give them a story worth repeating. Then organize that story until it becomes momentum. He did it in a pandemic. He did it in a campus election. Now he is doing it for software companies, and the play has a name: Imagine AI.
Yang is the co-founder and CEO of Imagine AI, a San Francisco company that went through Y Combinator's Fall 2025 batch. The pitch is deceptively simple. Founders build their companies out in the open on X, but they close deals on LinkedIn - and almost nobody runs LinkedIn well. Imagine AI builds a detailed voice profile for each leader on a team, then runs the whole content pipeline: posts, strategic comments, and a shared calendar so a CEO, a head of sales, and a head of marketing all push the same narrative at the same time. The reach compounds in a way a single founder posting alone never can.
It sounds like a marketing product. Yang talks about it like a physics problem.
The OriginA relief network run from a dorm
Before the startup, before the batch, there was Break the Outbreak. Yang started it in 2020, when he was a college student watching the pandemic outrun the institutions that were supposed to handle it. Most people that age were refreshing the news. Yang built an organization.
It grew to more than 400 people operating across 53 cities and 18 states. They distributed over 56,000 packages of personal protective equipment. The effort drew coverage from national outlets and commendations from a US Senator and a member of Congress. What is striking, in hindsight, is not the scale but the instinct. Faced with chaos, Yang's first move was to recruit, coordinate, and communicate - the exact three muscles a growth company runs on.
The CampaignElected to lead 32,000 people
At UC San Diego, Yang was double-majoring in Management Science and Political Science - a combination that reads, in retrospect, like a founder's origin story. He ran for student body president and won, becoming the youngest person ever elected to the role. The job put him in charge of representing 32,000 undergraduates.
He treated it as an operating role, not a ceremonial one. He lobbied in Washington DC and in Sacramento and helped secure $150 million in state funding for student housing. He pushed through wage increases for campus dining and housing workers. His campaign slogan doubled as a personal motto that he still repeats.
There is a smaller detail that captures him better than the funding figure. In the September before his term, Yang went skydiving. A person whose whole brand is "the sky's the limit" jumped out of a plane. It is the kind of on-the-nose gesture most people would be too self-conscious to make. Yang leaned in.
The CompanySelling the thing he was always good at
Imagine AI came together with co-founder Neo Lee, a former founding engineer at startups backed by MiraclePlus and Mysten Labs who researched game theory at UC Berkeley. Yang brought the go-to-market instinct; Lee brought the technical depth. Their shared observation was almost embarrassingly obvious once stated: founders broadcast publicly but do business privately, and the gap between the two is where growth leaks out.
The product is built around deep research. For every person on a client's team, Imagine AI assembles a persona report that can run past 100 pages, drawing on transcripts and internal communications to capture how someone actually thinks, writes, and sells. A voice engine the company calls Aura turns that research into drafts. Humans approve before anything ships. The claim is that it takes an executive's LinkedIn post from an hour of work down to about five minutes.
The proof point Yang likes best is that the company sells itself using its own product. Imagine AI was fully booked within its first week out of Y Combinator. On one call, a customer paid through Stripe inside fifteen minutes. That is not an accident of a good deck - it is content-market fit working exactly as advertised.
What clients report (aggregate, per company materials)
Figures reflect Imagine AI's own reported customer outcomes at launch. Bars scaled for readability, not to a shared axis.
The customer list skews credible: MongoDB, Rippling, Slash, Corgi, Paramify, Conduit. The backers do too - Y Combinator, Pelion Venture Partners, Founders Inc, Blast Club, and a roster of angels drawn from Palantir, Google, and Airbnb. For a company barely out of a batch, the signal density is unusual.
The ThesisContent-market fit before product-market fit
Yang's most provocative idea is that founders may be chasing the wrong fit first. Product-market fit gets the headlines. But if nobody reliably hears about the product, the product does not matter. Content-market fit - content that reaches the right buyer every single time - can be the difference between a good product that dies quietly and an average one that compounds. It is a debatable claim. It is also a very Sky Yang claim: distribution as a discipline, not an afterthought.
He extends the logic into a stranger future. Yang has floated a vision where buying and selling stop being human-to-human at all - where, as he puts it, your agents interact with each other instead of humans. It is a Westworld-flavored idea, and he says it with the same matter-of-factness he used to describe PPE logistics. Whether or not the timeline is right, the pattern holds: he is comfortable being early.
The ArcOne instinct, three arenas
- 2020Founds Break the Outbreak; scales it to 400+ people, 53 cities, 18 states, and 56,000+ PPE packages.
- 2022Elected the youngest-ever UC San Diego student body president, representing 32,000 students. Goes skydiving that September.
- 2022–23Helps secure $150M in state housing funding; serves as Co-Chair of the UC Council of Presidents.
- 2023Founds Svshi Inc. while finishing a double major in Management Science and Political Science.
- 2025Co-founds Imagine AI with Neo Lee; joins Y Combinator's F25 batch and books out in week one.
The tidy read on Yang is that he is a politician who learned to code a company. The truer read is that the categories were never separate for him. A relief network, a campus government, a growth startup - all three are the same task in different clothes: get people to believe something, then organize the belief until it moves. He has just kept raising the stakes on where he does it.
What he is selling now is not really software. It is the thing he was always good at, finally packaged so other people can rent it. And if content-market fit turns out to be as real as product-market fit, the kid who jumped out of a plane under a "sky's the limit" banner will have called it early - again.