While the exact article might have errors in them, I believe the general premise to be true. Lack of expected demand.
It faces multiple hurdles, like being a sidegrade compared to
IVB, pressures with abundant computing options, economic worries, an maybe the biggest of all: no one really needs to change their computer anymore.
They probably had to sacrifice some of peak transistor perf and architecture on Haswell generation and 22nm, to get better in low power segments. The only problem is, those segments aren't in their best shape right now.
Last year had what, 8 million Ultrabooks sold? That's a pitiful number considering significant number were likely replacement PC sales, while they lost greater volume in the overall market. They expect >50% growth this year, meaning maybe we'll see 16-18 million, if we are being optimistic.
You've got pretty attractive options in the Haswell Ultrabook category, but the price has rose correspondingly. Replacement models like the Vaio Duo costs $300 more than the last generation, when the last gen was >$1000 already!
Now the PC industry is hedging their bets on Windows 8.1, which pushes the ever so undesirable Metro interface, which will create delays in purchases and the same excuse as last year "8.1 will increase sales".