According to rumors I've heard, this joint venture is something their board wants because they're panicking over bad quarterly results, but the 18A process is actually coming along as more or less as expected, and the engineers don't see any sort of technology gap TSMC could fix. Their main issue was starting late. They've been unable to find foundry customers because running manufacturing as a service is a lot different than running it as part of a verticlaly integrated stack, and Intel is struggling to deal with customer expectations (this is a really common problem with any company that tries to turn an internal process into a service). The board wants a quick fix to all this and is flailing.