REGISTER | LOGIN
Breaking News
Blog

Apple and Google in the Memory Business?

NO RATINGS
Page 1 / 2 Next >
View Comments: Newest First | Oldest First | Threaded View
DMcCunney
User Rank
Author
Re: Flash is hardware
DMcCunney   4/13/2017 6:01:32 PM
NO RATINGS
@name99: Most optimal for Apple, I suspect, is having someone take over the memory fabs who is a lot more willing to be flexible on how they view themselves and what they're willing to do for partners. SO, for example, I could see the REAL story as being something like: Apple is willing to partner with Hon Hai (one way or another --- pay one third of the purchase price, sign a long-term contract, whatever) in this purchase, and Hon Hai commits to following Apple's R&D and manufacturing agenda.

I concur.  Apple is an enormous customer.  They don't need to own a fab, as long as the party that does puts Apple first in line for capacity, and will manufacture to Apple's specs with quality levels and pricing Apple finds acceptable.  I suspect Hon Hai would roll out the red carpet and say "Just tell us what you want. We'll make it."

The longer-term challenge for Hon Hai will be keeping up on process technology. (And I believe part of Toshiba's problem was that they hadn't.)  As process geometries shrink and new technologies are required. keeping a fab current can cost as much as building one in the first place.  In an interview here with TSMC's CEO a while back, he commented on that, and said something to the effect that only already having the fabs in place to upgrade made it possible, and he doubted he could start one from scratch now  because the costs were so high.  When a single piece of equipment used in a fab may cost half a billion dollars, you need very deep pockets.

>Dennis

name99
User Rank
Rookie
Re: Flash is hardware
name99   4/13/2017 5:09:36 PM
NO RATINGS
Suppose you're Apple and you have ideas for how improve your products using a new memory technology. For example

- using MRAM as a write buffer or cache for your flash OR

- using ReRAM as byte addressable single-address-space persistent storage

What are your options? Right now they're sub-optimal. You could buy separate chips from Everspin or wait for Intel to (eventually) get its act together with Optane and then sell you chips, but you don't have control over schedule, priorities, and manufacturing of exactly what you want on the same die, or at least on the same package.

Owning a "memory" fab gives you this control. Now it's not clear that it's the best solution. Apple's happy to let TSMC MANUFACTURE the SoCs it designs, and this works well in terms of letting others share the cost of improving TSMC's line.

Most optimal for Apple, I suspect, is having someone take over the memory fabs who is a lot more willing to be flexible on how they view themselves and what they're willing to do for partners. SO, for example, I could see the REAL story as being something like: Apple is willing to partner with Hon Hai (one way or another --- pay one third of the purchase price, sign a long-term contract, whatever) in this purchase, and Hon Hai commits to following Apple's R&D and manufacturing agenda. And then this nuanced deal gets garbled in transmission until it turns into "Apple considering buying Toshiba's memory business". 

 

Doug_S
User Rank
Author
Re: Leveraaging purchasing power?
Doug_S   4/11/2017 6:16:08 PM
NO RATINGS
The purchase of a fab is a capital investment, and doesn't act as a tax deduction. You will get one down the line as you depreciate that investment, but the money the EU says Apple owes Ireland is for past years anyway so anything that changes their current or future tax situation isn't going to affect that potential tax bill.

Don Herres
User Rank
Author
Leveraaging purchasing power?
Don Herres   4/11/2017 4:48:38 PM
NO RATINGS
Apple has lots of money parked in Ireland from overseas profits.  The US and EU are trying to tax it and the purchase of Toshiba's memory business would just allow them to invest before it can be vacuumed up in taxes.

Also, investing in Japan would be similar to investment in China.  Apple becomes a partly Japanese company which would improve their ability to grow in that market.

DMcCunney
User Rank
Author
Flash is hardware
DMcCunney   4/11/2017 3:36:35 PM
NO RATINGS
I'm amused by Google and Amazon being named as possible bidders for Toshiba's memory business.  As Jim Macgregor at Tirias Research pointed out, the memory business is cyclic.  I've been following the boom and bust cycle for decades, as something provides a huge demand for memory, producers jump in to fill the need, the need is filled, and producers late to the party are left with excess capacity and huge losses.

Flash is certainly in demand, and prices are dropping as capacity to make it grows.  A friend and his wife got 256GB Seagate SSDs as freebie swag at an industry trade show, which illustrates the dropping costs.

Flash is hardware, and neither Google nor Amazon are really hardware companies. 

There was an article on an Android oriented site a while back claiming Google ought to buy HTC.  My questions was "Why?  HTC makes the Google Pixel phone, but there is no benefit to Google in owning them.  The Pixel is a straight OEM deal.  They don't need to own HTC to get what they want."

Similar comments apply to Amazon.  It does sell the Kindle eBook viewer device and the multi-media Kindle Fire tablet, but the Amazon division responsible essentially oversees the OEM deals for the hardware.  Amazon isn't in the hardware business.  It buys from others.

The business model for a hardware vendor is completely different than the ones used by Google and Amazon, and it would be an error for them to try to become hardware vendors.  They can just buy what they want from hardware vendors, and are big enough customers to drive vendors to produce hardware to their specifications.  They have no need to do it themselves.

Who will get Toshiba's memory business is a fun question.  The joker in the deck is the Japanese government's desire to have a Japanese company acquire it, but the underlying economics make that unlikely.  The question is whether a Japanese owner could make money in the memory business, and the answer I can see is "They can't."  Pricing pressure is enormous, and I don't believe a Japanese firm can lower costs enough.  Korea and China are lower cost producers, and can.

>Dennis

 

 

Like Us on Facebook
EE Times on Twitter
EE Times Twitter Feed