Texas Instruments selling OMAP division
Laying down their ARMs
Jul 29, 2011 in Finance, Microprocessors, Opinion, Rumors
Word has reached SemiAccurate that Texas Instruments (NYSE:TXN) is hip-deep in the process of selling off their ARM/OMAP division. Given where the division is now, one of the few key players in the high end ARM space, now seems like a very good time to sell.
TI’s OMAP5 CPU is likely to be the first A15 based CPU out of the gate, and has semi-official Google blessings for the next tablet/phone/whatever is in this week for Android device, so the company is riding high. If you are going to pawn off a chunk of your business, now is about as good a time as any. Without being too specific, we are confident that the sale process is well under way, it isn’t just a rumor.
That brings up the biggest question, who is buying? The players that have been mentioned to SemiAccurate are Intel, AMD, ATIC, Nvidia, and a long list of other interested parties. We think any player with an ARM chip in process now would not be a good candidate, but some have pretty sparse roadmaps, so you never know.
Intel is an obvious one, but we consider this unlikely. Intel had, and sold, a top notch ARM division. Their new play in the space, Atom, is the biggest challenger to the ARM camp, and Intel buying OMAP would be an admission of defeat. In fact, Intel putting out any ARM chip at all, other than the ones they are currently making (Oh go look it up, Google works you know), would likely end Atom as a serious product. Don’t hold your breath on this happening.
The other seriously mentioned candidate is AMD, ATIC, or some combination of both. This one makes a bit of sense, especially if you buy the thought process that AMD is without a consumer electronics strategy, and needs an ARM chip. AMD had one. AMD sold one. AMD has a tablet chip. AMD does have a consumer electronics strategy, and if you understand it, you won’t ask nearly as many silly questions on the topic. Heck, AMD owns all three next gen console GPUs, so it’s hard to say they don’t understand the space.
The problem is, why would AMD want OMAP? AMD proper probably doesn’t, but ATIC could do quite a bit with the division, including getting them cutting edge process tech, and throwing money at them. Global Foundries could even license OMAP based IP to third, or would it be fourth at this point, parties. ATIC buying OMAP has a lot of synergies, and would only tangentially involve AMD too.
From there, the list gets very long. TI has a world class ARM line with their OMAP division, and consistently delivers good if not great products. That has a lot of value to many organizations, and phone companies like Nokia, a long time TI partner, HTC, and even Apple might be a good fit. There is no shortage of candidates, but the list right now is likely very short, if not one long.S|A
Updated August 19th, 2011 6pm
Texas Instruments contacted S|A today and provided the following statement:
“We’re aware of recent speculation in the press about the sale of our OMAP business, and want to set the record straight. To be clear, these are rumors, plain and simple. They are not true, and were not started by TI. TI remains committed to our core Wireless business, which encompasses the OMAP applications processors and wireless connectivity solutions. And, we are committed to helping our customers succeed in the marketplace.”
34 Responses to “Texas Instruments selling OMAP division”
Trackbacks/Pingbacks
-
-
Sep 15, 2011
[...] сообщает сайт SemiAccurate со ссылкой на собственные источники, компания Texas [...]
-
-
Aug 4, 2011
[...] Semi-Accurate has dug up a rumour that says Texas Instruments about to sell its ARM/OMAP division. [...]
-
-
Aug 4, 2011
[...] Semi-Accurate via Fudzilla .plistaWidgetHead { color: !important; background-color: !important; } .plistaItem [...]
-
-
Aug 3, 2011
[...] de un posible interés de Texas Instruments (tercer mayor fabricante de semiconductores mundial) de vender su división OMAP dedicada al diseño y fabricación de chips bajo arquitectura ARM para dispositivos de [...]
-
-
Aug 3, 2011
[...] news website Semi Accurate is reporting that Texas Instruments (TI) is planning on selling off their ARM/OMAP [...]
-
-
Aug 3, 2011
[...] rumor that Texas Instruments will sell off its ARM-based OMAP chip business. It was reported by SemiAccurate that TI may divest itself from the OMAP business and sell the unit to Advanced Technology [...]
-
-
Aug 2, 2011
[...] to ARMs: Texas Instruments Apparently Selling OMAP Division Tech news website Semi Accurate is reporting that Texas Instruments (TI) is planning on selling off their ARM/OMAP division. TI has not [...]
-
-
Aug 2, 2011
[...] news website Semi Accurate is reporting that Texas Instruments (TI) is planning on selling off their ARM/OMAP [...]
This is a bit surprising as Google is said to have chosen TI OMAP as the optimization platform for Ice Cream Sandwich (just as Honeycomb was optimized for Tegra2)
What makes Charlie think that AMD has scored all three next-gen consoles? I think they only have two (Microsoft and Nintendo). I have the feeling that Sony is going for something more exotic for the PS4 (Think what IMGTec bought a while back and they already supply the GPU for PSPVita).
This is the first victim of new R&D needed in ARM market to compete with nvidia that has started a new era.
Think about it: ARM market is booming, TI should have a good chance to get a lot of money out of it, but they need heavy R&D to create competitive products that can compete with nvidia, Qualcomm and Marvell ones.
Results: it’s better to sell ARM division to get most out of it while they show they are competitive and before they get sunk by rivals next generation.
Seen how much TI has invested during these years in the OMAP project (I saw a sample of the fist one when working for producing an handset) it’s surely a sign that the ARM market has changed and evolved towards more powerful/professional products than those created up to now.
Good for the market evolution, but a pity a player has left the game.
Oh, and I think LG should buy them – if they really want to enter the chip making game. But still, I think the end-fight will be between Nvidia, Qualcomm and Samsung. 3 is already too many, and there will probably be only 2 big ones in the future, and all the rest will fade out.
OMAP 5 won’t be the first Cortex A15 chip. Tegra 4 will be (and it will be more powerful, too – quad core vs dual core)
no Tegra 4 Logan
will not be be the first out with Cortex A15 level cpu
it will almost certainly be Krait that will be the first one, it will be shipping end of this year with devices likely early 2012.
and it will most likely followed by OMAP 5
and likely ST Ericsson A9600 after that if ST Ericsson can stick to the timeframe
one of the things that are different now is that nivida wasn’t one of the lead partners for cortex a15 like it was for the Cortex a9
Krait won’t use a Cortex A15 Core, it will use a Qualcomm inside core design, Cortex A15 compatible.
@ Marc
i did say “Cortex A15 level cpu”
No source, no story, worst reporting ever. Throw stuff on the wall and see what sticks then brag you are right 1 out of 10 times. And 1 out of 10 is being generous.
Awww pour pour baby Danni. Still upset Intel won’t pay you money.
Disappointing to see dismantled one of the top ARM developing teams. Let’s see what haves in mind their buyer.
maybe apple want to strengthen it’s designing office
after all I don’t think AMD go after ARM like intel they have their competing product
and like intel it look like defeat if they get an arm division .
by the way I guess admitting defeat in industry is not important what is important
is end of the year revenue . you have higher revenue than last year and nobody question
you even if you shut down 90% of the company .
arm means low power & small. most think up, anit x box.
hand held games toys, calculators with correct science & finance, dedicated stuff like xfm radio, astc devices, gps, even cell phones, less tablet, needs moore.
christmas gizmos,rpm meter or helicopter with loop ‘d loop. home security,power saver. vechile entertainment. anything aa battery operated thats multi functional. texas instrument makes small gadgets, credit cards, machine tools, hand held stuff.rotten cotton ball testers.
drashek texarkana
What an idiotic article! What exactly is TI selling? Every major component of their SoC is someone else’s IP. So ALL TI is selling is a bunch of engineers with certain industry experience- and the best of these individuals will be leaving as quickly as they can find new jobs.
Thus we have a standard US con job, and the sucker in these cases is almost always some Asian company (like Creative) hoping to level up. In the meantime, TI will rebuild another ARM division, taking every advantage of the fresh start, and the fact that better IP than that used by OMAP is flooding the ARM SoC market.
TI gets to offload an out-of-date team with out-of-date skills, and make a bundle doing so.
Google likes buying engineers… and from what I’ve heard of all the Android partners, Google “likes” working with TI the most.
Also, TI has a pretty decent reputation for working in the open source community with the linux kernel and well adopted platforms like beagleboard. There are good engineers at TI it seems..
Do you have links to your claims? Just curious, not wanting to start a flame war. I was under the impression that OMAP was one of the best.
Actually, OMAP chips have a decent amount of TI IP in them. Look at the IVA blocks, which have contained TI’s own C6x DSPs and fixed function video decode and image processing blocks. This IP is used all over the place in other TI SoCs and microcontrollers. In fact, TI sells devices under brands that aren’t OMAP but are very similar to existing OMAP products – check out DaVinci and Sitara for instance.
If TI is just selling OMAP that means that they’d have to be transitioning this IP from TI exclusive to third party licensable in some form. Or if they’re selling everything that includes this IP giving up a huge chunk of their product portfolio that is used in far more than consumer phones and tablets.
I wonder if the sale will include the fabs that build the chips.
When presented with list of possible buyers, I immediately focused on ATIC as most likely. Intel and AMD no, for reasons stated, GLOBALFOUNDRIES is not in the design business, but ATIC could always use a design bureau to go with GLOBALFOUNDRIES. For a while, it looked like ATIC would snap up all of AMD, but OMAP might actually work out better. Either way, more business for GLOBALFOUNDRIES.
Hmm, its not beyond GloFo to buy AMD and Omap and stick them together. AMD are a brand and it would make more sense to strengthen the AMD brand in the low power area than compete with AMD. Seeing as they own AMD and it would give AMD a good step into the really low power market then it would be a pretty good move for them.
Hell GloFo could pay for it and essentially “give” Omap to AMD in return for an increased share in AMD to the same value.
ATIC could buy it and start to create their own chip design company but, when they already own a reasonable sized chunk of AMD and IIRC are commited to buying another chunk of it in the nearish future, well starting from scratch just isn’t cost effective, if they really want to get into it, buy AMD.
Personally I think Glofo/atic could increase their stakes in AMD but can’t see it happening for a few years till GloFo have another couple of their fabs finished and in high volume production.
Given Google’s spending spree on IBM manufacturing patents, I think they would be a good candidate.
any links?
IBM Micro is still alive due to these patents and i don’t see them selling their best asset.
http://tech.slashdot.org/firehose.pl?op=view&type=story&sid=11/07/29/1257249
What about Chartered Semiconductor with their ARM license, I thought AMD/GloFo/ATIC was going to use that license if they wanted to make ARM chips?
A licence to use already prebaked ARM chips or use the ISA to build your own chip isn’t the same as buyer a different line of chips and with it an entire team dedicated to R&D with top notch ARM based chips.
Currently Glofo/AMD/Atic have entirely no rights nor access to anything TI do. You could make a 1000 different chips that are all based on ARM chips that will run ARM code, but aren’t remotely like each other and all with different performance/power/size/cost.
Having an ARM licence or chip is easy to get, having an end product, shipping, in products and having the next 2-3 chips in various states of readyness is entirely different.
GloFo already has a relationship with ARM holdings for manufacture of Cortex Cores.
By working with ARM, ARM has a specific process in mind to design their core while the partners who license ARM cores can enter into relationship with GloFo. The SOC manufacturers may decide to go to GloFo for SOC creation thus greatly cutting down the time to market.
http://www.globalfoundries.com/newsroom/2010/20100215.aspx
Charter Semiconductor is a part of Global Foundries after ATIC purchased it.
Unlike Charlie, I doubt if GloFo/ATIC will be a suitor as I do not see any advantage over the current situation vis-a-vis ARM designs.
With ARM wooing AMD in their Fusion developer summit, I think it will make more sense that AMD follows the Qualcomm model w.r.t ARM(license ISA alone but then design the core and SOC) rather than TI Model which licenses the Cortex cores. That way they can really differentiate from the other ARM SOC designers.
I might guess that AMD is waiting for a end date for the “non-compete” clause or something that it would be having with Qualcomm(which purchased all of AMD/ATI ‘s SOC related activities in 2008). This is a standard practice in Buyouts that you promise the buyer that for 5-7 years you will not enter the area that you had sold them.