The main argument they are using is AMDs Financials. They have referenced the losses AMD has had in 2012 and 2013.
There were a lot of unfortunate scenarios (management/staff 'issues'...not to mention ATi gents having a little too much sway post Ruiz) and missteps (The whole essence of bulldozer) that mixed with some things questionably beyond their immediate control (GF having a lot of problems and their associated contracts) that have held back a company that really does have all the pieces.
Rory (the CEO) is seemingly turning things around and attacking the right segments as they can, but sweeping changes don't happen overnight in microchips, nor will their seemingly bright future immediately be apparent to everyone (outside the folks like us that watch this stuff closely). A lot of it depends on things like process tech/memory (bandwidth) technology hurdles just not possible/feasible for them at this moment.
The turn started with the hiring of a bunch of ARM engineers (fairly immediately after RR took the reigns), as well as (and including) Jim Keller (whom is a one man game-changer). From there we have the ditching of Bulldozer for two archs that handle it's shared purpose in a more targeted manner, the cleanup at GF (that seems to have a straight path to 14nm products in 2016 thanks to Samsung), and the inherit benefits of doing so. A big part of their APU business will only come into it's own not if/when gpu compute starts to play a bigger role in applications, but also when we start hitting thresholds where their markets are satiated by their performance (and scaling.) I'm not going to say '4770/290x' for a desktop, but surely more competative ipc and mixed with something in the realm of 260x-like architecture will be upon us soon, and then perhaps a new arch with Tonga-like gpu by 14nm. Like-wise, when we start seeing something that is similar to the xbox360 (in performance) and then Kaveri's gpu in mobile, it will also be a big deal.
It also cannot be downplayed AMD set the yardstick for a gpu not only with the xbox360, but now with x1/ps4. To think they don't have a plan to make products (not only gpus, but tdp-targeted apus) that will not only be similar, but then exponentially scale with them is just asinine. That holds a lot of weight as we move through the next 6 years.
TLDR: Yes, their recent past rough. No, their immediate future isn't going to wow enthusiasts. That said, their targeted approach is working, and will likely continue to work as they not only enter more markets, but build their ecosystem through newer and better products, slowly climbing the 'good-enough' ladder for more and more people and products. Not only is the potential there for greatness in the next couple to few years, it's more-so on their doorstep than either Intel or nvidia, whom each lack a key component, and have their own design choice/process problems. AMD's time to shine is rolling around again, and all indications increasingly point that they will be ready to take the stage.
