Neither Amazon or Texas Instruments are talking about a report that originated in an Israeli financial paper that Amazon is in "advanced negotiations" to acquire all or part of TI's OMAP applications processor business.
Both TI and Amazon declined to comment, citing company policies against commenting on rumors. But a TI spokeswoman said she was misquoted in the original story, published by Calcalist, an Israeli business daily.
[Get a 10% discount on ARM TechCon 2012 conference passes by using promo code EDIT. Click here to learn about the show and register.]
The TI spokeswoman said she provided Calcalist only with the same statement she sent to EE Times, which cited TI’s policy not to comment on rumors and restated what TI already said about re-profiling its investment in OMAP to target embedded applications as opposed to the smartphone and tablet markets.
In making the announcement about re-profiling OMAP last month, Greg Delagi, TI's senior vice president for embedded processing, declined to answer questions about whether it meant TI would seek to sell a portion of the business. But Delagi made it clear that TI planned to continue developing OMAP in an effort to capture a bigger slice of the $18 billion per year embedded processor market.
It is unclear how TI could sell portions of the OMAP business that focus on smartphones and tablets while at the same time aim OMAP processors more specifically at embedded applications. "They would have to sell the whole division," said Will Strauss, principal analyst at Forward Concepts Inc.
I would agree. My initial reaction is that this doesn't make a great deal of sense for Amazon, but I do appreciate what Roger Kay says on page 2 of this story. Like everything else, it all depends on the purchase price. If they can get OMAP (or some portion of OMAP) for a song, maybe it makes sense. Otherwise, I don't know that it does. I am also not totally convinced that there really are advanced negotiations taking place on this.
Some years ago Nokia used to make their on SoCs and then suddenly decided to stop. At the time they annually sold about 400M phones and obviously reached the conclusion that it even at those volumes it did not make economic sense to keep designing asics. Also Nokia really excels at sourcing and logistics, that was never the problem.
From that I am quite convinced that there is very little reason for a phone maker to do custom SoC. The only reasonable explanation for Apple doing it is fear of getting cloned.
For Amazon to go that route is just ridiculous, they should drop Omap anyhow and look to MediaTek or Boxchip to cut costs even further.