http://ir.amd.com/phoenix.zhtml?c=74093&p=irol-EventDetails&EventId=5163539
can anyone parse these PDFs?
can anyone parse these PDFs?
Computing Solutions was down 20% YoY.
Yes, it really does look like the console deals (& Wall Street's addiction to junk bonds) saved AMD for now.
That's the only healthy business at AMD. They lost market share again to Intel, and given that the PC market is reacting they should have lost market share to Nvidia as well. This in a quarter where they launched new products, which should have generated good sales because of the inventory build up from OEMs. Next quarter, if their sales predictions are to be believed, they also will lose market share to Intel.
This quarter I won't even bother with the Q&A, they do not have anything new to show and nothing new to explain. We should see a lot of RR tap dancing and that's it.
They blamed bitcoin mining machine softness for some of the desktop MPU declines. Unbelievable.
Pretty typical. Doing well - we are great. Not doing well - it's not our fault.
They blamed bitcoin mining softness for some of the desktop MPU declines. Unbelievable.
Yes... I remember how annoying AMD Q&A can be when management starts to tap dancing.
Q: Rory, what's your CPU roadmap?
A: We plan a multi-prongued ambidextrous heterogeneous semi custom approach to our company, we are in the middle the transformation.
Q: Lisa, what's your plans for when Intel deploys 14nm chips.
A: We plan to use face recognition technology with our APUs in order to provide value to customers.
Yaaaawn
The 14-nm and 16-nm nodes from GlobalFoundries and TSMC are 20-nm FinFETs. 20-nm bulk then 20-nm FinFETs, same node generation different FETs. Then, around late 2017 / early 2018, AMD will move onto true 14-nm FinFETs called 10-nm FinFETs. There will be a split at the 7-nm node for the pure play foundries; Tunnel FETs, Carbon Nanotube(Big bias for this one), FinFETs, SOI FinFETs(GAA/Omega).They talked about the potential for further opex reductions during the next 12-24 months. I do not know how they can reconcile this with the claims that they will be moving to the 14/16nm FinFET nodes in 2016 (which is still fairly aggressive relative to the other fabless companies). I just don't see how they can afford to keep pace.
Perhaps I am missing something; would appreciate your thoughts on this.
Intel plans to do this with the 10-nm node with FinFETs and Carbon Nanotubes. ~10-nm FinFET to ~10-nm Carbon Nanotube, same node generation different FETs.
I wish you were joking, but you're not.
They talked about the potential for further opex reductions during the next 12-24 months. I do not know how they can reconcile this with the claims that they will be moving to the 14/16nm FinFET nodes in 2016 (which is still fairly aggressive relative to the other fabless companies). I just don't see how they can afford to keep pace.
Perhaps I am missing something; would appreciate your thoughts on this.
Intel is doing pre-emptive tooling at Ireland and Arizona for Carbon Nanotubes + EUV + 1.5D BEOL. So, there will be a slow down in 14-nm and 10-nm FinFETs and 10-nm CNFETs will be early.Where has Intel said ANYTHING about production chips being based on Carbon Nanotubes?
The shift to 16nm finfet isn't too far fetched if they go straight to 16nm, skipping 20nm on their MPU and only moving their TSMC GPUs to this node. Given that GLF 20nm is a late mess with effectively 0 customers adopting it, I don't think that skipping it is such a bad deal.
I'd agree, but AMD has publicly announced ARM/small core X86 on 20nm.
If I had to guess, they may "launch" 16 FinFET K12 in 2016 in the same way that Temash "launched" in 2013.
Intel is doing pre-emptive tooling at Ireland and Arizona for Carbon Nanotubes + EUV + 1.5D BEOL. So, there will be a slow down in 14-nm and 10-nm FinFETs and 10-nm CNFETs will be early.
Where has Intel said ANYTHING about production chips being based on Carbon Nanotubes?
Source? I'd be really interested in reading more about that.