Canonical's Leann Ogasawara started off the kernel session by basically saying 3.13 is what more or less will happen, but then other developers chimed in that made Linux 3.14 a greater possibility. Given that Linux 3.14 is not likely to arrive until March or April, it's stable release is too close to the April debut of Ubuntu 14.04 LTS to put much trust in its quality and fear of regressions. As a result, Linux 3.13 is the safe bet.
However, Linux 3.14 might end up happening depending upon how the Linux 3.13 kernel looks. Canonical is concerned about the hardware enablement support in Linux 3.13 and if everything landed. It sounds like their main concern is the hardware enablement of Intel Broadwell as the i915 driver was brought up.
Broadwell made it into 3.13, but it's not fully-baked. The Intel developers themselves have said the 3.13 support is experimental and will take more time / kernel releases to iron out the support.
Canonical is willing to back-port a great deal of code from 3.14 and other kernels if needed, which would result in yet another "franken kernel" for Ubuntu Linux. This is most likely what will happen since when 3.13 stable is released the Broadwell support and other new hardware enablement will still be in a less than desirable form for end-users; it happened with Haswell, Ivy Bridge, etc, in taking more time than was estimated to have reliable Linux graphics support. It appears that they wouldn't back-port the entire 3.14 DRM tree though because the i915 driver was specifically called out during the vUDS kernel session since it's notorious for regressing on older generations of hardware.
We'll see what happens but Ubuntu 14.04 LTS will likely be on a Linux 3.13 kernel with a ton of back-ports over not having faith in Linux 3.14 shipping timely and in a ready form that's dependable and low on regressions. Hitting Linux 3.13 is at least exciting though over being ultra-conservative and sticking to Linux 3.12 and doing greater back-ports because Linux 3.13 has a ton of great features.