Chartered Semi and Closing ThoughtsChartered Semi
Recently ATIC, the parent company of GlobalFoundries, announced a bid to purchase the foundry company Chartered Semiconductor, and all signs are pointing to a successful completion of this acquisition in the near future. Assuming that GlobalFoundries and Chartered Semi will in some fashion be merged together into one foundry (it only makes sense), this offers up some incredible benefits for the company as whole.
First, Chartered has customers, lots of them, while GF only has one to date. Once the merger is complete GF will essentially have dozens of customers and moving them between the current technology found at Chartered and the higher tech options at GlobalFoundries Fab 1 and Fab 2 should be a simple undertaking. Customers will like the ability to combine orders for low wattages 130nm chips with high performance 28nm wafers to a single customer.
No. Words. Needed. Nerd alert.
For the GF team the Chartered fabs offer up yet another opportunity for “re-use” of the company’s expensive R&D. Technologies that are proven to be successful and profitable can then be slowly siphoned down to the less advanced fabrication plants that Chartered owns – I could even envision a physical re-use of manufacturing tools – like a $2 million version of hand-me-downs.
Immediate benefits will include the ability to offer process technologies higher than 65nm to customers that GF would not otherwise have been able to do. To put a fine point on it: this combined GF will be able to offer just about every type of chip technology to just about anyone.
The Future of GlobalFoundries
While we are still in the early stages of GF as a non-exclusive foundry, the future is looking pretty solid for them. It will take about 12 months from the official signing of ST Micro (GF’s first non-AMD customer) before product will be produced and put to market and the same can be said for the first of the ARM licensee’s (which have yet to be announced). The high performance ARM Cortex A9 processors will be built on the 28nm High-K metal gate offering from GF and the design infrastructure should be ready for it by the second half of 2010. I will be curious to see if this CPU is made anywhere else (since it is licensed by ARM’s customers, they will have that option of course) so we can get a performance comparison directly between GF and TSMC.
Another interesting note made by GF’s team last week was they were “not interested in owning IP” but only wanted to build it. This is obviously a jab at Intel who only recently decided to start offering some of its Atom core IP for production at TSMC for third-parties (a revolutionary move from Intel though) yet will still of course produce the same architecture for themselves at their own facilities. The insinuation here is that Intel will definitely be able to product the Atom IP-based CPUs better than any other fab facility and any customer that chooses to use TSMC for a similar CPU would be at a huge disadvantage in terms of both power and performance. GlobalFoundries wants nothing to do with these kinds of conflicts of interest it would seem.
GlobalFoundries still has a long way to go to prove themselves to be a competitor to TSMC in the real world; right now we are only basing our opinions on a single announced customer and no products available in the market other than AMD processors. Even so, it is hard to deny that many fab-less companies will find the goods that GF is offering very attractive – including ATI and NVIDIA – and because of that the foundry world is already in turmoil. For customers, both of the corporate and consumer variety, the additional competition can only be good as it promotes the R&D and technological advancement we love. How different would the 40nm generation of GPUs have been with GF around a couple of years earlier? It’s hard to say for sure but I like to think these types of product shortages would be a non issue at the very least.
Just one of many things to look forward to with GlobalFoundries in the picture.