AMD won't die, they just may have to bow out gracefully form the super-high-end enthusiast arena. I see problems with the amount they can pump into R&D, and don't forget that whereas before they could pump all of their budget into researching CPU technology, the acquisition of ATI has meant that they now have to split their R&D and development costs over two seperate fields, CPU's/chipsets, and GPU's.
On the GPU front compare this with Nvidia who can focus all or the vast majority of their resources on GPU's and thus devleop/change new and faster architectures much quicker. Then on the CPU/chipset front compare Intel with their comparatively limitless resources who can pump obscene amounts into developing their CPU and chipset tech, and thus keep ahead of AMD by a considerable margin, at least now and in the future now that they have an incredibly strong architecture to work on. And the upcoming Nehalem? Scarily, scarily quick by all accounts, I genuinely can't see AMD having the resources to counter it.
In summary I don't think AMD will ever recover properly in either sector, their resources are now too thin, and they are making a thumping loss, the fact they've just had to sell 8% of the company and filed an SEC report showing huge deficits is a very telling sign of the position they are in now. I see Intel/NV sewing up the expensive enthusiast/high-end market, and AMD concentrating on the mid/value segment and making a profit.
their position hasn't changed greatly in a decade, they've always been in debt, yet generated large and increasing turnover. to compete in these kinds of battles requires large investment which means massive debt at first. but while the debts are massive, the company has a MUCH bigger future potential earnings than well, most companies around. intels turn over is immense, and in the future amd can hope to be making similar amounts. they will go through another 5-10 years at least of massive debt and expansion, but at some point get to a level where say they have 10 fabs, make 10 times the chips they do now. but when they hit that level they get rid of a lot of problems. for instance, switching to 65nm for intel is a case of shutting one of many fabs down, throwing money at it, and having it up and working quickly, then doing another one, all the while with production dropping maybe 5-10% for a month or two. for amd with 2-3 main fabs, that would be closer to a third drop in output for a while. they've hit their hardest period right now, getting contracts with dell means a minimum number of chips a year really, which in turn means that downtime is even harder on AMD.
they couldn't afford, literally, to change all fabs to 65nm, only to change them over to barcelona shortly afterwards. lower chips per waifer but guarenteed quota filling isn't ideal, but means they keep contracts. more chips per waifer, but having 2-3 months where they might be so short on chips that they could even risk losing contracts with the likes of Dell is simply not doable.
once they get to a certain level, size, and production capability those problems go away, completely. i suspect amd will sell more shares over the next 2-3 years, and get a lot more capacity somewhere/how, a deal with tsmc seems fairly likely at some point in the future too.
but the fact is, if they can get another 15% market share, considering the market will increase a lot in the next decade as more countries in the world get to western levels of tech levels and computer use, that will be massive.
as for investment in R&D, aquiring ati makes no difference really. in general both companies separately would input a certain amount of profits into R&D, those profits are the same with companies together or apart. however apart not only chip design, but materials, silicon, manufacturing and future ideas for making chips are all basically the same research being done by two companies, with them working together those costs are cut. intel do the same research on their own, as do ibm, and tsmc and all the other guys that make chips. they can also help each other with little tweaks and fixes they find and share tech which will help cut research costs quite a bit.
the main problem with AMD is people not understanding the debt is largely not a problem, because of future projected earnings, banks get this. if say, ati on their own had that debt, banks would have foreclosed years ago. banks and investors keeping AMD going means they understand and fully support AMD, if they were dead they would have been out long ago.
Its just a very difficult period, with winning massive fights with Dell they did great, but its a large responsibility to keep those contracts, which will mean 2-3 VERY tough years for AMD.
