Electronics Weekly Electronics Design & Components Tech News
NXP has said it can identify $500 million in long-term cost savings and $200 million in cost savings in the first 12 months after the merger. $200 million equates to several thousand lay-offs and $500 million to 10,000+ lay-offs. NXP says the projected cost savings don’t all come from lay-offs but this is where the bulk of cost savings usually come from in these cases.
Freescale has 17,000 employees but, as these areas of overlap suggest, some of the post-merger sackings may take place at NXP.
Here, according to OWK, are the areas of overlap:
ARM processors will be merged keeping the i.MX and QorIQ ranges but scrapping the rest.
Power Processors will be ‘de-emphasised’ although PowerQUICC will be kept for now
Freescale’s Kinetis competes directly with NXP’s LPC range so Kinetis will be chopped but LPC will of course survive.
Freescale DSPs will be fine – here the NXP products may be killed off
Analogue and Power Management will be merged – lots of good stuff from both which will be merged but only needs one sales force
RF – Freescale devices are optimised for US and NXP for Europe so less risk. These users won’t be willing to approve alternatives and will expect all devices to continue. Indeed any attempt to scrap parts would cause action from US Govt.
Sensors – here Freescale has the better devices, NXP ones could gradually die off
Wireless – total overlap. They both do everything, even down to Zigbee and RF4CE. But these are all dominated by software and users won’t want to port. If they kill anything the users will just port to TI !! A really hard one this.
If NXP can sell off the RF power business to help secure this deal it shows that even good business units are at risk, Dan, this is being done for money not for product development.
Freescale has set Kinetis apart through their software investment. The Codewarrior IDE is one of the best on the market currently and MQX has been gaining in the RTOS market by leaps and bounds. They’ve also built an amazing ecosystem around their Kinetis development kits. It would be a huge shame if it was cut.
Freescale Kinetis is a very successful product, so why do you say it “will be chopped” ? I realize NXP also has a Cortex M4 and I don’t know which is more popular, but isn’t the NXP part a funky big-little arrangement that costs more than Kinetis?
Well I hope it works, but it still seems a strange deal to me. NXP seemed to be beginning to be on the recovery track and had some good products and had cut the internal crap that had existed for ages…. I would be more concerned about Freescale’s survival, but then NXP would be mad to buy them and take on their debt unless there was a good reason (or this was a PE idea).
Yes Mike it seemed opaque and unconvincing to me.
“He pegged the “cost synergy”—the savings in operating costs expected after the two companies are merged—to be about $1.2 billion in R&D. The savings will be mostly coming from supporting various CAD tools and models out of one organisation than two, Clemmer explained.”
I really do NOT think that either of these companies CAD strategies is so bad that there is a saving of even $100m to be had. If your company has more R&D employees then CMS want a larger fee for company wide licences.
Interesting Agnolus, but a little light, or I should say a lot light, on the prospective lay offs.
did you look at this:
http://www.eetasia.com/ART_8800710354_480500_NT_19bdf2e3.HTM?click_from=8800122481,9950150066,2015-03-05,EEOL,ARTICLE_ALERT
Product divisions that don’t appear to belong to NXP’s stated focus on “security and connectivity with a smarter world” are Freescale’s network processors and NXP’s standard products, … Clemmer, however, did not deny the possibility of letting the two divisions go, “if they can be sold at their full value.”
AND
NXP, in announcing the deal, stated that it will sell its own high-performance RF unit in order to avoid regulatory issues
Well, ARM is already releasing far more advanced Cores and On-Chip Fabric than FSL at this point. PPC has some legacy, but ARM is already ahead of FSL PPC from uArch POV.
FSL’s remaining expertise – may still have an edge:
Embedded Flash for Auto,
Hardened Std Cell for Auto,
control blocks for Auto motors,
Analog components for Auto,
& Auto SoC system architecture .
Due to lack of Key Capital Investment from the previous buyout, FSL is in no position to compete at higher end.
Low cost design and product commoditization is now apparent. Even low power, design automation is so wide spread – most design firms can do now …
Massive layoff will be inevitable sooner or later 🙁 .
I do not agree with rIoT – LPC and Kinetis parts cannot simply stay untouched even in relatively short term perspective. From business point of view such redundancy is pointless cost-wise. Generic industrial market has parts with relatively short lifespan – this is not aerospace or auto – and that enables them to merge MCU families and trim ‘excess’ fat.
Regarding PPC MCUs in automotive powertrain, if they get rid of it not only will they scrap their historical roots and stronghold in powertrain, but also (and more importantly) completely destroy their differentiation in this market. Once you go with ARM, you find yourself competing only with peripherals and manufacturing yield. Then it becomes mass competitor market with low margins – typically this is devastating for most supplying companies. Also, trying to stitch eTPU timer with ARM may not be trivial… Fortunately their customers may protest and help them there – it’s not that you can easily accept decades long investments and know-how in powertrain PPC being thrown away just like that.
There sure is – which of ST or Renesas is the main and which is the second source.
Foot – aim – fire !!
Ha Ha, Mike, clearly there are going to be some challenging choices to be made.
just spoke to an RF expert friend of mine who makes commercial avionics.
Me : “Do you use Freescale power devices ?”
Him : “Yes. Very good devices”
Me : “Do you have a second source for them ?”
Him : “Of course. It’s a FAA and CAA condition of approval”
Me : “Who’s the second source ?”
Him : “Oh NXP. I’ve got one on the rig now to make sure it matches the Freescale one we ship.”
Me : “Not a second source anymore. NXP just bought Freescale”
Him : (after long pause) “S*** !!!”
Interesting riot, and this is going to get more interesting. As you say they won’t want to move too quickly but the deal has been predicated on $200m of cost savings in the first 12 months which indicates product pruning involving several thousand job cuts in Year 1. Cost savings of $500m ‘long-term’ suggest 10k+ job cuts and some substantial effects on product lines. My suspicion would be that shareholders’ interests will prevail over customers’ interests and certainly over employees’ interests. While a $9.5bn debt for companies which are not very profitable makes their product investment plans very vulnerable in the event of rate rises.
David, while I agree on some of the ARM part – specifically keeping the i.MX and newer QorIQ parts, I can’t see them scrapping Kinetis or LPC ranges – they are too valuable for them. I’m sure there will have to be some alignment, and trimming over time, but if you do this too fast you end up like Renesas (pre-NEC) with lots of old product lines you have to maintain for key customers (such as LPC in the Apple cables).
I expect they will fold together the MCU teams with a new brand moving forward, but have to maintain the existing product ranges for a very very long time…..
Automotive may be the big “Win” here for ARM. Traditionally powertrain is PPC based (or Infineon Tricore) – if through the merger they look to finally put down PPC this could be a big socket win for ARM Cortex-R processors.