As of yesterday, partners can no longer participate in the first $500,000 a startup raises, unless it's three weeks past Demo Day.
This is because, YC founder Paul Graham writes, outside partners look at YC partner-funded startups as an indicator of how promising it is, which is bad for the YC startups that don't receive investments from its partners.
"[...] as there came to be more YC partners and they paid more attention to picking likely winners, till by the last couple batches, other investors could treat investment by YC partners as an accurate sign of how promising we thought a startup was," Graham writes. "Which meant we were now making it harder for the startups that partners didn't invest in to raise money."
Here's the full post from Graham:
"As of this batch we're introducing a new policy for investments by YC partners in the companies we fund.
YC partners have invested in the startups we fund since the first batch. In the beginning it was harmless, and occasionally even necessary. And other investors couldn't treat investments by YC partners as much of a signal when I was the only one doing it, because I was so haphazard about it. But over the years this gradually changed, as there came to be more YC partners and they paid more attention to picking likely winners, till by the last couple batches, other investors could treat investment by YC partners as an accurate sign of how promising we thought a startup was. Which meant we were now making it harder for the startups that partners didn't invest in to raise money.
Our new policy is designed to prevent this by depriving other investors of this signal till it's too late. The new rule is that YC partners can't be in the first $500k a company raises, unless it's 3 weeks past Demo Day. And since a startup's fundraising trajectory is almost always established, one way or another, by 3 weeks after Demo Day, this should mean that we can't affect anyone's fundraising unless they've raised $500k, at which point their fundraising is already successful.
This should fix the problem. If it doesn't we'll try something else."