Aditya Agarwal’s anti-incubator incubator
How South Park Commons became a top incubator by bucking the deadline-driven, burn-and-churn model of most other incubators.
Written by Evan Armstrong
Illustrations by Jesse Auersalo
The siren song of startups, the call to build something new, perhaps to create a company ex nihilo, is one that often compels the ambitious. Maybe you yourself have felt such an urge to go and build your own thing. However, knowing you want to begin building is a lot easier than knowing where to begin.
In 2015 in San Francisco, around ten people who were all experiencing this feeling gathered to meet in Aditya Agarwal's home. They were discussing everything from governance to startups to artificial intelligence (AI), but they had one thing in common: They were eager to figure out what they wanted to do next.
“I would come home from Dropbox after a long day of work, and there would be people in our living room or kitchen hanging out and discussing AI,” Agarwal remembers.
The meeting had been organized by Agarwal’s wife, Ruchi Sanghvi. In 2012, Sanghvi and Agarwal had sold collaboration software startup Cove to Dropbox and joined the company; Agarwal had become the CTO of Dropbox and Sanghvi the VP of Operations. But now, both Sanghvi and Agarwal were ready for something new.
These few simple ingredients — a group of open-minded, curious people, the desire to start something new, and two experienced leaders — helped create the base of South Park Commons (SPC), which Agarwal makes clear to me is Sanghvi's baby.
Today, SPC has evolved from being a space for free ideas to being one of the top startup incubators in the world for founders, if measured solely by how many successful startups it has produced. It has resulted in the creation of over 130 startups, with 13+ unicorns in the investment portfolio, giving it a total valuation north of $35B. Some hits include Web3 darling Compound and SaaS giant Airtable. SPC has raised two funds. And the first $55M fund has already been returned with additional liquidity coming down the pipeline.
But none of this was visible from SPC’s humble origins at Agarwal and Sanghvi's place in 2015. In those days, they met a few times a week to talk freely about the topics that interested them — early versions of the whiteboarding sessions that happen today with SPC team members, guest speakers, and technical paper discussion groups.
The conversations went on long. Meetups were chaotic. Members debated between a variety of governance models including elected bodies, top-down leadership, and even anarchy. Eventually, they settled into a more centralized leadership structure, with Agarwal and Sanghvi appointed as leaders of the community.
Eventually, Agarwal explains, there came a point where he became so desperate to go to bed at a normal time that he and Sanghvi paid for an office space in South Park, San Francisco, formalizing the community for the first time.
“Life is too short to bear this cross that you have to work on something really important, but you're not having fun while you do it," says Agarwal
“It got to the point where I was like, ‘Ruchi, I cannot come home to all these people in my kitchen…I'm going to kick you guys out. This is crazy. I cannot have people in my house every night talking until 3:00 AM about this shit."
They also decided on a goal. Rather than just focusing strictly on startups, being a place for co-founder dating, or helping companies go from 0 to 1 (or even 1-2 like, say, Y Combinator), they decided that SPC must focus on taking people from “-1 to 0.”
“We are not trying to create a vibe where the first question anyone will ask you is some variation of: ‘Hey, which investors have you talked to? Hey, how much money are you looking to raise? Hey, what's your next milestone? Hey, what are you doing for demo day?’
The vibe is more like, ‘Hey, what are you building? What are you trying to explore? Is there something I can help with?’”
This is one of the paradoxes of SPC that is initially hard to grok. The incubator’s lack of focus on startup success is exactly why its startups are successful. Agrawal calls their model the “anti-incubator incubator.” In comparison with the deadline-driven, churn-and-burn model of pumping more and more companies that some incubators follow, SPC tries to do something it believes is much grander.
A community member “graduates” once they have chosen what is next after SPC or if their company idea grows big enough. As of 2022, there are roughly 150 people in SPC at any single time. It has about a 10% acceptance rate and a tenure that is flexible depending on what a person wants from SPC.
Some people have gone on to start companies — over 130 and counting since SPC's start — but there is no strict obligation to do so. That’s another key difference from other incubators. No one is forced down any path. Some members have run for public office, published a book, joined think tanks, or even become monks (seriously).
SPC optimizes its process mostly for technical talent (meaning engineers or coders of some kind) but is also open to anyone with grit and horsepower. Members have even included former Olympians.
Originally the companies that were coming out of SPC raised capital from outside investors. Sanghvi and Agarwal, who have 130 angel investments under their belt, would often invest themselves in startups. But there was no fund.
To formalize and expand the organization, the husband and wife duo raised a $55M Fund 1 in 2018 and a follow up fund of $140M 2021. Now, SPC has a nearly 10 person full-time team split across investing and community management.
Agarwal explains to me that playing the startup game requires being comfortable with very long odds. When successful outcomes only occur for ~1% of startups, it takes a remarkable amount of luck and skill to build a career. The number of people who have been early employees at multi-billion companies is very, very small.
Agarwal knows this well — he was one of those people. He was not only one of the first engineering hires at Facebook, but also became the CTO of Dropbox as it scaled and went public.
When I ask him about what his secret is, on how he was able to create a career by building and betting on outliers, I expect a grand soliloquy or some novel framework. Instead, his answer is simple:
“You should obviously consider the mission and the impact, and the kind of work that you'd be doing. But a huge part of it is simple — is it fun? And do you enjoy the people?”
I’m surprised, so I raise my eyebrows. But Agarwal has an answer to my incredulity.
“Life is too short to bear this cross of having to work on something that's really important, but you're not having fun while you do it,” he says. “I just don't think that people do their best work if they're not having fun.”
This ethos of people having fun while they work started in 2005, during the earliest days of Agarwal's career. Agarwal, a 22-year-old software engineer looking for his first job in startups, was introduced to a Harvard dropout named Mark Zuckerberg.
Agarwal was invited to the Facebook headquarters (then a house in Palo Alto), where the primary decorations were beer bottles and pizza boxes.
Engineers were strewn around the house, hacking away and writing code. Agarwal found himself attracted to the company, but it wasn’t because of its code or product — it was because everyone looked like they were having fun.
He officially decided to join Facebook, but Agarwal gives credit to his then-future-wife Ruchi, who he had been dating since college. Sanghvi insisted that the product was going to be something big and also planned to join the company; when she did join Facebook, she was one of the first female engineers, worked on the News Feed product, and was the product lead for Facebook Platform.
Note: This author highly recommends finding yourself a spouse who can convince you to take a job at one of the most successful technology companies of all time.
Over the next five years, Agarwal scaled Facebook's engineering while helping to build Search, Infrastructure, Newsfeed, and a variety of other products. Most people would be satisfied with scaling a product from 500K users to 800M — not Agarwal. Saying in an interview that he left “mostly because [he] wasn’t learning quickly enough,” he departed Facebook to cofound Cove with his wife. They sold it to Dropbox.
Eventually, the same feeling emerged — Agarwal wanted to build something new. This time, lucky for Agarwal, there was a gang of interlopers in his kitchen having similar feelings.
The startup problem doesn’t just stop at idea generation and validation. There’s also the fact that most people can’t afford to spend 12 months without earning any salary unless they have money from some previous success. It can be uncomfortable to say “I’m just exploring” for 6-12 months when people ask you what you are up to. Some folks end up doing something suboptimal like hacking on an idea on nights and weekends. Perhaps what is most detrimental for society is that many lack the resources or creative space to start and end up continuing on in corporate America, full of regrets that they didn’t build something better.
To that end, SPC started a fellowship program last year that gives $400K to teams looking to explore startup ideas. SPC is a selective program: The first 4 cohorts had a less than 1% acceptance rate with ~50+ teams already going through the program. Some cohorts have had a general focus, while others have had industry focuses; one of the more recent groups focused on Web3 and they are considering focusing next on biotech.
The community is also expanding. SPC recently opened a second office in New York. “Covid threw us for a loop,” says Agarwal. “Historically, we’ve had a strong in-person culture, with a lot of serendipity. But I think what Covid did for us at SPC is help us figure out the right hybrid mixture. Can we scale using a hybrid mixture? Should we start new offices? I believe there is a need for an SPC in every tech hub in the world. Seattle, Boulder, Austin, Miam, London, Lagos, Bangalore, Mumbai, Singapore, Beijing.”
With all of these procedures, stats, offices, and alumni, things might start to feel an awful lot like a structured process with ascribed outcomes. However, Agarwal refuses to let that be the case — he wants SPC to be something more than that.
“We do not want to be bound by the fucking rules of traditional venture capital,” he says. “The last thing we want to be thought of as is a venture capital firm.”
When I asked him what the future should hold, the sense I get is more communities, more cities, more startups. And most importantly, more fun.
“We have to continue evolving. That's what makes this job fun for me. I don't want to be doing the same VC job 10 years from now. That's not fun at all. In 10 years, I want to be figuring out what it is that I need to provide to allow people to build cool new things. That’s what gets me up every day.”
More Like This
Packy McCormick, one man show
By Meghna Rao
Gaby Goldberg, online to on-chain
By Meghna Rao
Boris Wertz on web1 and web3
By Luke Thompson
How Jonathan Hillis’ cabin became a DAO
By Shreeda Segan