broadcom to close

[tinytim]

http://www.theregister.co.uk/2014/07/23/chips_are_down_at_broadcom_thousands_laid_off/

not sure if this has been posted - i used the search function and couldnt find anything relevant so i guess not.

[dannyf]

Not unexpected.

Two trends are very unfavorable to guys like TI/Broadcom/Marvell:

1) chips are becoming more and more highly integrated: You find more and more functionalities on Qualcomm chips for example.

2) the handset market is more and more bifurcated: you have iPhone / Samsung on the high end, and MTK/Huawei on the low end, with their own chipsets. TI made the decision to exit earlier and couldn't find a buyer for their product portfolio. I see the same fate for Broadcom / Marvell.

[Noise Floor]

Concur with Danny, however I have noted the Broadcom has been investing heavily in baseband technology.  Might be a few years but they are making a play for it.

[free_electron]

Now now.. Broadcom is not closing. They are stopping the development of baseband processors for cell phones. This is only like 5% of what broadcomm does...

For those not in the know : the baseband processor is the chip that sends data (voice, sms, internet data) over the mobile network. It is a digital processor that handles the transport. All you need is a radio transmitter and receiver) the baseband does the encoding decoding, error correction , channel hopping, and much more. Even modulation and demodulation is done in digital domain.

Baseband processors are very time consuming to develop and you have a tremendous battle at hand since Qualcomm basically sits on all the good stuff (read: the patents... ) so you need workarounds for their techniques. This costs time and money and you end up with a more power hungry product.

And the there is the other problem : basebands are a commodity. You can only make money off the really fast ones with the latest functionality if you can sell those to smartphone makers. The old skool phones don't need the hi-end baseband processors so there is no money there.
As for the hi-end phones .. Who you gonna sell to ? There is only apple and samsung. The rest are insignificant in terms of volume.

So yeah. The market is too small and cannot support four or five baseband makers. Two is sufficient.

[gxti]

Don't worry, THE INTERNET OF THINGS will save them.

[Stonent]

I guess everyone will have to buy a Beaglebone black then.

[SirNick]

All set!  I got a spare just in case.

[ju1ce]

Broadcom is actually doing quite well, it's just the mobile technology unit (majority of which was bought last year from Renesas, who bought it from Nokia in 2010) that's not profitable enough. I think that this is no big news except for in Finland where lots of old Nokia engineers go out of a job, especially in Oulu where Microsoft is closing down its site as well.

[bluejonah]

I don't think they're closing. They will probably merge with another company in the future, however.

[SL4P]

Don't worry, THE INTERNET OF THINGS will save them.

... and THE CLOUD!

[Stonent]

Send in.... The clouds....

[rob77]

i don't think broadcom has to worry about the future... they're doing well in the network segment and also in CNA segment (converged network architecture).

[SirNick]

They're certainly prolific in the network space.  Although, if I had a choice between a Broadcom server NIC or an Intel equivalent, I wouldn't choose the Broadcom.  Same goes for consumer WiFi chipsets.

I'm not a rabid OSS zealot, but as a hobbyist, their policy regarding SoC (ala RasPi) openness is mildly frustrating.  As compared to the BeagleBone Black, I appreciate how TI jumps into the water with the hobbyists and hackers -- even if they do still require you to use closed-source blobs for some of the stuff where they have significant IP.  The board layout, schematic, source code -- everything else -- is available for the taking, with their blessing.  It's refreshing.  (Granted, the CPU on the BBB is not a media player designed to decode a ton of patented formats, so that might have something to do with it.)