The development cycle for kernel 4.15 was a clusterf*#k. On the other hand, kernel 4.16 is expected to have a smooth development cycle.
Simply put, given a choice, Ubuntu 18.04 LTS should be based on a kernel having a smooth development cycle, and not based on a kernel having a clusterf*#k of a development cycle.
Yes, that’s exactly what I’m afraid that they will do. Backports don’t come without risks. What’s more, the backports will need to be maintained for five years (to support those not on HWE).
So, given that 4.16 will have been released before the Ubuntu kernel freeze (April 12th), it would make absolutely no sense to release Ubuntu with 4.15 and then backport the Meltdown and Spectre fixes from 4.16 into 4.15.
The Kernel Team can track the kernel release candidates. By about 4.16-rc5 (expected mid-March), the rc kernel should be fairly stable, which should provide the Kernel Team enough time to work out most wrinkles (and apply the fixes to final 4.16) before the release of Ubuntu 18.04.
But even if they miss a few bugs before the release of Ubuntu, no big deal. The Kernel Team always misses a few bugs before an Ubuntu release. That is, whether they go with 4.15 or 4.16, bugs will need to be fixed after the release. Better to fix them in 4.16 (smooth) than to fix them in 4.15 (clusterf*#k).
Of course they can. But as you point out, this will be an LTS for Ubuntu. An LTS release shouldn’t be based on kernel that had one of the most chaotic development cycles in recent Linux history.
Kernel 4.15 will have lots of baggage and should be avoided. Rather than carry that baggage for the next five years, better to wait for 4.16. Or if waiting isn’t possible, then I suggest releasing Ubuntu 18.04 LTS with kernel 4.14, which, although also needing Meltdown and Spectre fixes from 4.16, at least will receive them as a part of LTS kernel support.