Earlier quoted context omitted.
ARM's profit last year was 170 million pound last year, that $68 million per quarter vs intel's $3.2 billion. There are of course the big chip manufacturers ,but also they don't make much money. Qualcomm made around $200 million per quarter. It would take a huge amount of growth for this markets to become financially meaningful to intel. probably not gonna happen. As for future growth , i think there could be plenty…
I didn't realize ARM was so much less profitable than intel. It's partly because ARM's model of development is shared with its licensees (e.g. Apple's A5 is based on ARM, but customized inhouse, and I believe uses third-party components); so a fairer comparison of the ARM "business" would include all these designers, assemblers and distributors, as well as the chip manufacturers you mention. The ARM model is "dis-int…
And regarding the critical code sections becoming silicon: i don't think it's becoming silicon , because of the huge volumes that are require to manufacture silicon. but they might become FPGA code.
FPGA is basically a programmable hardware. it has been used in the past as computer co-processor , in fields like biology, oil and gas and database accelerators. but the main barrier is that programming them is complex and it's hard to find FPGA programmers.
But i see improvements in these areas. convey computer claims that it has made programming an FPGA as easy is c programming. it woud be interesting to have convey computer offer via the cloud.
But let's return to the intel vs ARM servers: according to arstechnica , arm would find it very difficult to offer high per-thread performance at low power. but given a cheap enough arm server processor , and the help of GPU's(and FPGA or other coprocessors) intel could be seriously threatened.