Amazon looks to acquire TI mobile chip business, report says | Hacker News Reader

Amazon looks to acquire TI mobile chip business, report says

news.cnet.com

68 pointsdrone24 comments

Rewording of Engadget's rewording of The Next Web's translation of

http://translate.google.com/translate?sl=auto&tl=en&...

We need more efficient ways to get tech news than via chains of rewordings, indeed.

This post's title is inaccurate and needs fixing. The article's title is: "Amazon looks to acquire TI mobile chip business, report says".

It's an external report which is claiming that the two companies are in talks. It's not a press release from Amazon or TI saying they're doing it.

[BTW: what is flagging meant to be used for? Why is there no documentation on these features?]

Flagging is the extreme version of a downvote. Rather than expressing the uselessness of a post, it expresses the need for the post to be reviewed by a moderator (removed from the site altogether, reworded headline, merged with another article). Generally it's used on articles that are obvious troll articles, completely illegitimate articles, or highly inaccurate headlines.

Wonder if this means no more Beagleboards/pandaboards, etc...

Thinking the same, but encouraging a hobbyist / developer effort around the OMAP has huge strategic benefits for Amazon.

I think the worst decision they could make would be to isolate development by killing off the nascent BB market.

I would think it's in Amazon's overall interest, not necessarily in Amazon.com's interest, to further ingratiate themselves in the developers community.

That being said, much of the TI development kits are sold at-cost, or even nearly below (see the launchpad series, $5 development boards), so it may not be as economically feasible for Amazon to match that pricing, giving that selling bulk chips aren't their business.

Launchpads would not be affected by this deal which only is about the OMAP series.

I know, I was implying that beagleboard/pandabaord may be being sold at a limited profit much like the launchpad boards, to attract interest to the platform.

Yup

This is awesome. I'll have a cheaper ARM-based EC2 host please.

Has anyone considered that this could also be for ARM-based servers instead of just Kindles? :-)

I could be wrong, but I think this is only talking about TI's mobile baseband chips, not ARM processors.

If they are referring to TI's OMAP processors, then they typically feature multiple ARM cores with various media accelerators (http://www.ti.com/lit/ml/swpt034b/swpt034b.pdf). These are like all-in-one chips designed for mobile platforms that are optimized for displaying media on a LCD-ish display at minimum power. The OMAP offerings are pretty similar to chips from Broadcom or Samsung, however, it makes sense for them to get integrated with their biggest customer. They already have NDAs in place and TI probably has field engineers that essentially work for Amazon, they would just be transferring financial stake different production processes.

Having said that, the reason that you probably don't want to run a EC2 instance on an OMAP, or any ARM for that matter, is because they are not 64bit compatible, they do not have good virtualization support, and you would not be taking advantage of various media/display capability.

Well, it's a very poorly sourced article, but TI has been trying to unload its old baseband mobile chip division so it could focus on OMAP 4.

They bought the company, not a specific type of chip. They can do whatever they want and steer this in any direction they want.

arggg really dude? I dont want my web site on that please!

Is there anything inherently wrong with ARM servers that give you legitimate reason to immediately balk at the suggestion? As far as I can tell, ARM servers are no less reliable or speedy as x86 servers. They just cost less to own and run. In fact, there's already a sizable precedence of ARM server clusters like EC2 (and including EC2). See:

[1] http://www.ubuntu.com/download/arm

[2] http://www.h-online.com/open/news/item/Canonical-releases-EC...

[3] http://arstechnica.com/information-technology/2012/07/your-v...

[4] http://www.phoronix.com/scan.php?page=news_item&px=MTA5N...

Yes this argument is becoming true with time because ARM is really gaining phase. But we have Virtualization and its already cheap and works best at the moment. I dont see why anyone should go for separate ARM server ( in normal scenarios of course ) instead of using virtualization

Considering how many Kindles Amazon sells at break even or at cost to them, this would indeed be a good business for them even if they kept all of TI's output to themselves. SoC has got to be an insignificant part of the cost with margins etc.

Actually the processor itself is normally fairly cheap, the LCD is normally the largest price (or storage, depending on what type and how much space is provided.)

There are enough companies making ARM processors that I'm not worried about the loss of competition; Barnes & Noble could easily switch to NVidia or Qualcomm or Samsung or...

This would be a good move for Amazon. Look at the biggest phone/tablet sellers right now - both Samsung and Apple design their own chips and exercise a great deal of control on the manufacturing as well. At a certain scale, there is a big benefit to making your own hardware. If nothing else, it lets you optimize for your own requirements instead of for the requirements of other customers.

I think it could be a good move. Execution will be everything here, and it is a fairly new market for them (anybody know who designed the Kindles?)

Busines is mispelled in the title.

Reply on news.ycombinator.com