jennyh :
Nobody said it was a free lunch.
GF have a bunch of customers, and a bunch of fabs. Each customer pays money per wafer, which in turn helps GF to make a profit etc.
What profit? Global Floundries isn't a publicly listed company, yet they have taken over loss making enterprises like Chartered and AMD's Fabs.
Where is the evidence that these rocks tied together can float?
GF are already #2, having taken over UMC on revenues. How can that be so? Simple, AMD are already paying GF for wafers, and the Chartered revenues too.
Revenue != Profit.
Now AMD may be getting a preferential rate, but in no way will the removal of that preferential rate cost them more than the drain of running the fabs did.
We actually don't know if this is how it will play out yet.
Obviously that is the medium term goal, but if AMD aren't able to significantly increase it's market and profit share, then it won't be cost effective for Global to keep subsidising AMD. It is still far from clear if Global Floundries will be able to achieve the economies of scale that allow them to charge AMD a price that allows them to compete against Intel.
Let's say for example that currently each CPU AMD makes costs them $20 in manufacturing costs and each CPU Intel makes costs them $10.
What happens if Global Floundries need to one day charge AMD, cost + profit for GF, thus have to charge AMD $25 per CPU and Intel's costs are still $10.
That in and of itself is not a game changer, but it would show that AMD's goal of competing against Intel just got that bit much harder than it used to be.
If it did, they wouldn't sell.
Now here is where you show your lack of understanding of the circumstances.
AMD sold because without selling, bankruptcy would have been guaranteed at some point in the medium term(maybe it would have been in the short term), but by selling their fabs to ATIC, they buy themselves probably another 10 years to try and find some way to stay alive.
But whereas it might have seemed possible in the heady days of K8 dominance that AMD could become the equal of Intel, the selling off of their fabs now makes that seem an impossibility.
GF *can* run many fabs profitably, because they have many customers, just like TSMC.
Not all companies are equal.
GF is a group of enterprises that have been bleeding red ink, it is a tad optimistic to assume that by tying these rocks together they will now be able to float.
The difference here is, GF needs AMD because AMD will be the company who pushes them further. GF needs AMD to be successful, and AMD will always be the #1 customer.
This is absolutely true in the short term to medium term, but unless ATIC are complete idiots, they must have done due diligence and worked out if it was a realistic goal that they could be a true rival to a group like TSMC(who don't fab any AMD cpu's), over the longer term and thus be able to do it without AMD if need be.
In summing up, I believe you have confused the medium term with the long term.
If ATIC/GF were so convinced that AMD would be absolutely crucial to them in the long term, and if they are prepared to bankroll AMD for as much as it takes, why didn't they bankroll AMD so that they could continue with their court action against Intel,(especially as you believe that Intel's guilt was so obvious), and go for a massive damages payout, instead of settling for a measly $1.25billion?
You have been suggesting that ATIC/GF may be able to afford FABS for 16nm production at a cost of $50billion each and Intel won't be able to, yet ATIC/GF wouldn't bankroll AMD here(and the Intel amount shows that $1.25billion was all they needed to bankroll AMD for)
When you work out the answer to that question, you might then realise how fanciful all your predictions of the AMD/GF partnership being one that will sweep all before it.