Since BFL isn't able to deliver, why doesn't AMD do a SHA256 ASIC line?

[bitcoinnix]

It seems like ATI could design, fab, and deliver easily and in a short timeframe, given their existing infrastructure.

[1715302029]

1715302029

[1715302029]

1715302029

[tycho]

It seems like ATI could design, fab, and deliver easily and in a short timeframe, given their existing infrastructure.

Because the market is tiny compared to the markets that they're already targeting.

[bitcoinnix]

It seems like ATI could design, fab, and deliver easily and in a short timeframe, given their existing infrastructure.

Because the market is tiny compared to the markets that they're already targeting.

So's the Linux market but they're dumping lots of money into that right now on hopes of continued growth.  Seems the potential for growth would be well worth it for SHA-256 and sCrypt mining.

[Foxpup]

It seems like ATI could design, fab, and deliver easily and in a short timeframe, given their existing infrastructure.

Because the market is tiny compared to the markets that they're already targeting.

So's the Linux market but they're dumping lots of money into that right now on hopes of continued growth.

Say what? The Linux market is huge. The majority of servers and virtually all supercomputers run Linux as well as a significant number of embedded devices. The bitcoin mining market is absolutely microscopic by comparison. But don't worry. I'm pretty confident AMD will develop bitcoin mining ASICs once the market's big enough.

[bitcoinnix]

It seems like ATI could design, fab, and deliver easily and in a short timeframe, given their existing infrastructure.

Because the market is tiny compared to the markets that they're already targeting.

So's the Linux market but they're dumping lots of money into that right now on hopes of continued growth.

Say what? The Linux market is huge. The majority of servers and virtually all supercomputers run Linux as well as a significant number of embedded devices.

I'm a Linux sysadmin by trade so I get the market size for servers - but I can promise you that almost nobody is using AMD proprietary drivers on their Linux servers (It's a pain, even with DKMS & akmod to deal with the changes that kernel updates bring to modules).

I more meant that the Linux high-performance desktop market is microscopic compared to the rest of their line, so I'd bet bitcoin ASICs (or really even just really nicely done SHA-256 & sCrypt FPGAs) would be worth it to them.