The upcoming Y Combinator class will have somewhere between 80 and 95 companies, which will be the biggest in the program’s history. One class hit 83 companies, and then Y combinator pulled back, because, as founder Paul Graham wrote at the time, “We grew too fast.”

We’re still a bit mystified about what happened. Why was 66 [startups] ok and 84 [startups] not? Is there some kind of hard limit somewhere between those two numbers? Or will we be able to morph YC to get past that bottleneck as we always have in the past?

The companies in the upcoming class will average valuations of $10 million, and their value will likely be more than $1 billion when they graduate, Altman says. More than 6,000 companies apply each year. Altman was appointed president earlier this year when Graham, who founded and ran the program, stepped down.

Y Combinator has gone through a series of changes in the last year. The program switched its investment structure, ending a vehicle YCVC, which invested $80,000 in each startup from its twice-per-year classes. As before, Y Combinator itself will invest its standard $14,000 to $20,000 per founder in each company from an internal fund that has no outside LP’s. But the follow-on funding of $120,000 for each startup will come from an independent fund, which has no name. Fortune reported on this new fund in March.

MORE: Y Combinator’s YC VC may lose the actual VCs

Altman said on stage today that the point of the new fund was to eliminate any signaling issues when the venture funds in YCVC, who invested automatically in every Y Combinator startup, but did not necessarily make a follow-on investment each company’s Series A round. Those conflicts and signaling issues “got worse and worse every with batch,” Altman said.

The new structure also allows Y Combinator to get a bigger chunk of equity at a more reasonable valuation. “Valuations have moved up a lot, and the amount we get diluted [after] the initial investment has gone down. So we moved our terms a little bit too,” Altman said.

Altman wouldn’t comment on the size of Y Combinator’s new, unnamed fund, but said, “It’ll last for years and years.”

The total value of all companies that have gone through Y Combinator since it was founded in 2005 is between $20 billion and $30 billion, which is much greater than any of the accelerator program’s competitors. With the rise of Y Combinator and its two most well-known success stories, Airbnb and Dropbox, many copycat programs have sprung up. Few have produced breakout hits, though.

Altman says it’s a huge negative when startups have already gone through another accelerator program. “You should have already been accelerated,” he said. Through prodding by Michael Arrington, who interviewed him, Altman admitted that Y Combinator has a monopoly on startup accelerators. “So you’ve sucked 99.9% of the oxygen out of the accelerator [category],” Arrington asked. “Unlike other stages of investing, there is a monopoly.”

“We get the best companies, which is 90%,” Altman said.

Note: This story has been updated to note that Altman’s comment about selling for hundreds of millions of dollars was in reference to Hacker News, not Y Combinator itself, and that the company gets 6000 applicants per year, not per batch. (There are two batches per year.)