After 10 years of teaching startups to find product-market fit then grow like crazy, Y Combinator is finally following its own playbook. 

When the famous Silicon Valley accelerator began in 2005, it was just that: a prestigious three month program known for pumping out 100 companies every batch. 

It has spawned some of the current tech boom's biggest names, including Airbnb (worth over $25 billion as of its last round), Dropbox (worth $10 billion), Heroku (which Salesforce bought for over $200 million), and Reddit.

Yet, over the course of the past year, Y Combinator has transformed far beyond its roots. Sam Altman, who took over as CEO from founder Paul Graham in February 2014 and was recently promoted to president, has led its transformation. Under his leadership, it's launched a VC arm, a fellowship program, and a research lab. 

It's all part of Altman's plan to make Y Combinator 10 times bigger than it is today.

"The batch size had doubled since I started. We were talking about how do we add another zero, go from 200 companies to 2,000," Altman said. 

The first 10x

Y Combinator has done it before. 

When Y Combinator started in 2005, founders Paul Graham, Jessica Livingston, Robert Morris, and Trevor Blackwell incubated the first batch in Cambridge, Massachusetts, before moving the program to Silicon Valley. Among the first companies was the social network Reddit, and Altman's own location-based mobile company Loopt. 

There were eight companies spawned from that batch.

The three-month incubator now produces around 100 new ones, twice a year

“Obviously we have not scaled in the way Facebook does, but the program now is rather unlike what was here in 2006, both in number of startups, but also the breadth of companies, the quality, and the amount we deliver," said Paul Buchheit, the creator of Gmail, who joined Y Combinator in 2010 and now runs the incubator, recently renamed Y Combinator Core.

To Graham, one of Y Combinator's founders, there's never been a limit. As for as how big it could grow, even a number that seems surprisingly high likely isn't that much of a reach. 

"The number of startups is increasing.  I've always thought Y Combinator would have to grow with it. Otherwise you have to turn down applicants you think will succeed. It's very hard to predict how big you'd have to grow to avoid having to do that," Graham told Business Insider in an email.

A fund to tackle bigger problems

For its first 10 years, Y Combinator (the startup version of itself) was fine-tuning its product. It leveraged its network of alumni through its own social network, aptly called "BookFace." It applied software along every step of the way, including to solve the problem of how to get investors to commit to funding its companies.

Y Combinator's first act came to an end when Graham passed the baton to Altman in 2014. 

From the beginning, Altman wanted to bring more hardware and biotech companies to Y Combinator. These are capital-intensive ventures, like re-imagining nuclear energy or solving cancer through new testing -; not your typical growth hacking or VC-subsidized coupon for a free massage on demand. 

He soon learned their needs were different. The timeline is much longer and more capital intensive. A social app can go viral overnight, but changing America's energy resources takes a different approach. 

For nearly five months, Altman talked to potential investors (limited partners) and tried to figure out how to support these kinds of companies. He ruled out a debt vehicle, or a fund that would invest in early stage companies or non-Y Combinator companies. 

Instead, Y Combinator decided to expand its scope with a new VC fund that would focus on later stage companies that had graduated from the original incubator program. It was abold move, putting Y Combinator into more direct competition with traditional VC firms who often invested in its companies. 

By April 2015, the Continuity Fund, as it would be named, started to take shape, but it was still missing a leader. 

Altman didn't want a traditional late stage venture capitalist. Instead he looked at the the Y Combinator community and picked Ali Rowghani, who had joined the company in 2014 as a part-time partner after stints as chief operating officer and CFO at Twitter and CFO at Pixar.

"It kind of came out of the blue a little bit. It wasn’t something we had been talking about or something that we had been planning on, or even at that point something that I wanted to do," Rowghani told Business Insider.

Working with startups had been a fun diversion for Rowghani as he pondered his next career moves. 

"I didn’t join Y Combinator one day to one day lead their growth investing," he admits. 

When Altman asked him over sushi, he was still hesitant. But Altman convinced him by telling him his work had been exactly what these startups needed as they moved into later states. Plus, he could use his operating knowledge to help these companies grow to a larger company, like Twitter or Pixar.

"By that point, I was just so impressed by Y Combinator as a platform. I just saw it as so unique, so differentiated," Rowghani said.

Splitting the program to deal with growing pains

There was another issue to tackle. Y Combinator had become a victim of its own popularity. 

The incubator had seen demand double in the fall of 2014. Altman taught a class at Stanford, which reached entrepreneurs far outside the hallowed campus halls. 

"Of all the single things we’ve done for recruitment, that class has been the best," Altman said. "Our application rate basically doubled after the class."

However, Y Combinator's accelerator program demanded a lot from founders. They had to move for three months to the Mountain View campus in the middle of Silicon Valley. For founders with an idea but no money, sitting in some place like rural Pakistan, there had to be a better option. 

"Let’s say we have 600 founders in a year, and there’s 7 billion people in the world. I think we’re reaching just a tiny fraction of the talent that’s out there," Buchheit said. 

But he adds, "your brain would melt" if a demo day--the showcase where startups present ideas to investors--lasted two weeks instead of two days to cover increased number of companies.

At the same time, Y Combinator was worried that it was missing out on the next Facebook. The main accelerator had started to see more companies that were already functioning, instead of the "two people with an idea" companies that characterized its early years.

"That’s something that we worry about because historically companies like Facebook or Microsoft start off in a pretty primitive state. Obviously we’re worried about missing those companies," Buchheit says.

To discuss next steps, Y Combinator's leadership gathered at a campsite in the redwoods, two and a half hours north of San Francisco. 

The solution: A separate early-stage fellowship that didn't require the same in-person commitment as the original program.

A month after the camping trip, Y Combinator announced its inaugural fellowship class, led by Kevin Hale. The shorter eight week program required founders to devote themselves entirely to proofing their ideas and find a product market fit. 

The original program, now called Y Combinator Core, continues under the leadership of Paul Buchheit. Now, Altman is focused on the bigger picture--and for the first time since he took over, advising startups in the class.

Altman is watching the split carefully. Some founders have been confused if they should apply to Y Combinator Fellowship or the Core. The leadership decided to change the applications to avoid confusion, and Altman has already noticed a lot fewer emails as entrepreneurs begin to understand who they should be reaching out to.

"As long as we can keep the mean quality per company improving all the time, then more companies is better," Altman said.

One more arm: a research program

Even after all that, Altman believed there was more Y Combinator could do to help the startup ecosystem, and that's how Altman ended up sitting on stage with Elon Musk at the Vanity Fair Summit in San Francisco last October introducing a new project: Y Combinator Research.

Its first research nonprofit is Open AI, or researching how AI can grow without presenting a threat. Another research area that the group wants to tackle is basic income

It's a pet project for sure of Altman's and one that he's now devoting half a day to every week.

"We always view of our mission at Y Combinator as enabling innovation. And mostly that’s meant funding startups. but if there’s something outside of that and it fits our mission, then we will still do it," Altman said. 

A week after Y Combinator Research was announced, the Continuity Fund formally launched. The name Y Combinator had gone from meaning one accelerator (with a side project running Hacker News) to five different entities.

Y Combinator had reached its Alphabet moment. But unlike Google, which changed its name to Alphabet when it upheaved its corporate structure, Altman decided to keep the name. 

"I thought about it briefly. I like the name Y Combinator. I think it’s meaning is very good for what we do, and I think it still fits that meaning. In the way Y Combinator is a function that starts other functions, we are a company that starts other companies," Altman said.

This story first appeared on Business Insider