RHEL6

It is possible that Red Hat thinks supporting four releases at once was too much, and has decided, but not yet announced, that it will release RHEL less frequently, but with long phase 1 (new hardware and features) support. This might be a good choice because as the product matures, Microsoft have seen that customers become more and more resistant to upgrading. Rightly or wrongly they think it's an added cost (not a software cost, both Red Hat and Microsoft offer products where you don't pay for the specific version you use - but a training and testing cost) for no value.

So whereas you might have found customers would start using RHEL 4 on some production systems six months after it came out, and were using RHEL 5 within a year, Red Hat may have data that suggests if RHEL 6 was available today, it wouldn't drive any sales until 2011 anyway. So why spend the extra money on engineering? Wait until customers are begging for the new features, right?

Fedora offers them an advantage here. If Microsoft has some trouble that leaves them without even a beta of the new Windows Server for an extra year, they've got no way to put (server) features out there for customers to see. With Fedora, Red Hat gets an opportunity to show the way every six months, which must be good for customer confidence, while at the same time the rapid turnover of Fedora releases minimises ongoing engineering support overhead & makes it worthless for the roles where you might plausibly sell an RHEL license.