8. Intel has announced is revising its licensing agreement with AMD since the new deal apparently violates its original terms. However this is likely to be favorably resolved because can help Intel negotiate a resolution for AMD’s ongoing antitrust lawsuit.
9. The transaction may have to go back to NYS legislative bodies for approval, surely after the election and under a very different public finances reality in light of Wall Street’s collapse.
10. The transaction has reduced the downside risk for AMD and piled up significant risk on The Foundry Co. Clearly The Foundry Co is compounding risk factors making it a far riskier proposition than having relocated what is left of AMD. Things can go wrong for The Foundry Co either through: (1) Insufficient additional demand for over AMD's future capacity utilization; or if (2) AMD fails to secure/retain a sizable market share of the global microprocessor market, thus implicitly making The Foundry Co less viable.
The Foundry Co is exciting news for NYS in the sense that NY has now purchased a ticket in the global semiconductor lottery by becoming AMD’s and The Foundry Co. venture partner. NYS will share some of the rewards provided that The Foundry Co timing is correct amidst a complex and unusually uncertain execution plan.
AMD’s announcement about its breakup in two companies, recapitalization and commitment to Luther Forrest are all promising deals for NY, much uncertainty remains. A few observations are in order to moderate the hype of the announcement:
1. AMD is not coming to NY; the newly formed The Foundry Co is coming. AMD remains in Austin and Silicon Valley.
2. The Foundry Co is a relatively well funded startup, having a hugely complex business model, operating in at least three time zones (GMT+2, GMT-5, GMT-8) and carrying substantial assets and debt.
3. The Foundry Co has only one customer at the moment, AMD. The success of The Foundry Co rests on its ability to operate at full capacity. This is not simple because the global demand for leading edge logic processes is very limited with ~70% of it satisfied internally by Intel.
4. AMD which will have to utilize, at least for several years; most of the capacity of Fabs 36 and 38 in Dresden; and then of Fab 4X in NY. Fab 36 will continue to meet the demand for all current and near future AMD chips. AMD graphics chips, currently being produced at TSMC will migrate over to Fab 38 in Dresden which will be tooled in a rush.
5. Fab 4X utilization plans remain very unclear due to AMDs very small MPU market share of recent quarters. Supplementing AMD’s fab demand, here is the possibility that IBM might decide to shift some of its gaming console business to The Foundry Co. Finally, The Foundry Co. may also find ways to compete against TSMC for the IC cell phone business, still a growing market.
6. The Foundry Co competes against TSMC that has a 20 year lead on them, owns about 70% of the global foundry business, and operates in a single time zone (Taiwan) Fabless Semi demand today can be satisfied with just a handful of MegaFabs. (>> 100,000 Wafer Start/Month), and very little is leading edge. This implies that around 2012 The Foundry Co Fab 4X (32 nm) in Malta will be competing head on against the largest Fab of TSMC (28 nm).
7. A far bigger issue is The Foundry Co. capital expenditure budget announced to fall between $3.6 and $6 billion for the next 4 years (2009-2013). AMD spent over the past 4 years $6.4 billion (2004-2008), and resulted in a disappointing performance. For comparison, Intel is expected to have a CAPEX of $30 billion (2008-2013). Even more confusing is the fact that The Foundry Co will be now a startup company having a larger number of fabs to support with a much smaller proportional budget. We must assume divestitures, massive debt buildup and additional capitalization; none of that possible without first re-writing Intel’s cross licensing agreement.
Posted on October 9 at 5:25 p.m. (Suggest removal)
(Continued from previous comment)
8. Intel has announced is revising its licensing agreement with AMD since the new deal apparently violates its original terms. However this is likely to be favorably resolved because can help Intel negotiate a resolution for AMD’s ongoing antitrust lawsuit.
9. The transaction may have to go back to NYS legislative bodies for approval, surely after the election and under a very different public finances reality in light of Wall Street’s collapse.
10. The transaction has reduced the downside risk for AMD and piled up significant risk on The Foundry Co. Clearly The Foundry Co is compounding risk factors making it a far riskier proposition than having relocated what is left of AMD. Things can go wrong for The Foundry Co either through: (1) Insufficient additional demand for over AMD's future capacity utilization; or if (2) AMD fails to secure/retain a sizable market share of the global microprocessor market, thus implicitly making The Foundry Co less viable.
The Foundry Co is exciting news for NYS in the sense that NY has now purchased a ticket in the global semiconductor lottery by becoming AMD’s and The Foundry Co. venture partner. NYS will share some of the rewards provided that The Foundry Co timing is correct amidst a complex and unusually uncertain execution plan.
Clearly, it is still too early for celebration.
On Editorial: Time to accept AMD