Forbes magazine dismisses Qualcomm's defeat in the
Why does Qualcomm keep losing in the courts? It's lost in a German court, a
Now we know it can't be the latter, because Qualcomm CEO Paul Jacobs said recently: "We've not been a litigious company."
So does it pick the wrong fights? That may be because so many of the lawsuits are about patents which the courts find to be irrelevant or unenforceable.
So does it have shitty lawyers? Well its lawyers in one case were reported to the California Bar Association for gross misconduct, but that may have been more to do with them failing to get relevant information out of Qualcomm than anything they deliberately did to mislead the court. Anyway tthe US is full of lawyers and if Qualcomm can't find some smart ones it must be very dumb.
So why does Qualcomm keep losing? I think it's because they're trying to do the impossible. With 80 per cent of their revenue derived from royalties on licensing IP, and only 20 per cent from chip sales, they want to establish legal domination over the wireless industry which would force every company to pay a fee on their products to Qualcomm on pain of being debarred from the
When the worldwide industry adopted CDMA at 3G, Qualcomm seemed to see its chance and started demanding its fee. After that, the worldwide industry will never willingly adopt a Qualcomm standard again.
There are just too many powerful players in the world wireless industry for any one company to establish such a legal domination.
That worries Qualcomm because it thinks its royalty-based, licensing business, is going to shrink.
In the long run, that would probably be good for Qualcomm because it would oblige them to sell damn good chips for a living, just like it used to do in the days when it was simply a very much-admired pioneer of wireless chip technology.
Hey David, as far as I know Qualcomm's UMB is basically dead. I know someone who was interviewed for a job that was supposedly for both UMB and LTE a few months ago and he told me that they said they were focusing nearly 100% on LTE. And recent public reports indicate as much really.
The big question for Qualcomm at this point is whether they'll manage to grab enough royalties for LTE and/or WiMax to remain highly profitable. So from a business perspective, this will obviously be very challenging and a legal failure on that front would be a disaster of epic proportions.
From a technical (chip) perspective, it doesn't really matter though: right now Qualcomm has 3 fixed-function baseband blocks: GSM/EDGE, WCDMA/HSDPA, and CDMA/EV-DO (with the first obviously being the smallest by far). This is combined with an ARM9 shared by all blocks. LTE will be a distinct fixed-function block and that's about it really; UMB would have been another one, and they can mix-and-match everything as they see fit for different markets.
However the most distinct fixed-function blocks there are, the more their disadvantage will grow versus software-centric solutions such as Icera's Livanto and NXP's EVP16. At the same time, neither their application processing capabilities nor their connectivity technology is anything to be overly proud of; many other companies compete or beat them just fine in those areas.
In many ways, Qualcomm's situation is quite similar to Intel's right now. It could turn out alright, but they also have huge vulnerabilities in a lot of areas and their growth opportunities are not very credible. All IMO...
--Arun
P.S.: I've lurking your blog for some time, nice stuff! :)
Arun, thanks a lot, your comments are much appreciated, and what you say about the challenge Qualcomm's faces in asserting an LTE/Wimax IP position is most interesting, especially the parallel with Intel. I just hope no one gets to dominate the wireless industry, best wishes, David
Well IMO, the problem with the PC industry is that traditionally there has only been one value center, the CPU. Everything else was a commodity. Intel's problem is that this value center is likely to shrink in value and unit volumes won't increase fast enough to compensate.
Similarly, mobile phones used to only have one value center, the wide-area baseband (where Qualcomm used to lead). Everything else was a commodity. Similarly that's changing now and once again pricing is likely to decline faster than unit volumes are increasing, forcing diversification.
When there's only one value center, it's easy to have a monopoly. But it's much harder to control an entire industry if you are only a small part of the pie. Qualcomm has had a typical Intel-like reaction here (although they've moved a lot faster!) by integrating application processing and Bluetooth a few years ago. Now I think they have been forced to become a bit more moderate in approach.
By the way, I'm sorry if this is completely unrelated but... I've been very curious about Icera for some time and can't help but wonder: did Simon Knowles hint at any timeframe for their 40nm baseband during Wireless 2.0? Also, you mentioned in your article that they have a full CDMA license. Really? Like, say, VIA Telecom? Because if so they've certainly kept it very close to their chest and that's a very intriguing development if they exploit it properly! :)
No. Simon Knowles didn't give any indications on the timeframe for 40nm. On the Qualcomm license, I asked Simon and Steve Alpress very specifically for the extent of the CDMA license and was told that they can make a chip for any socket which Qualcomm can address.
I have heard on the grapevine, though no one at Icera has confirmed this to me, that Qualcomm did come back to them to re-negotiate the terms of the licence and, so I was told though it's not confirmed, initiated legal action against Icera, but Qualcomm (as usual!) lost the battle and Icera is legally entitled to to do anything with CDMA which Qualcomm can do with CDMA.
I agree this is a very valuable position to be in, and may help explain why Icera has been able to raise $140m and expects to raise another $60m this year when break-even is still 2-3 years away.
Very intriguing indeed! Hopefully they also have the information on how to implement it though and it's not like the situation NVIDIA had with QuickPath (they had a license covering all sockets Intel could possibly make, but it didn't explicitly say Intel would tell them how to implement it... you can guess the rest). Although in this case at least, there's probably enough info out there for them to do it on their own.
As for how much money they raised - I think it's hard to overemphasize how much potential they have, so while I'm sure CDMA might help, just think about how attractive a (speculative) roadmap like this would be:
Mid-2008: 65nm Baseband, 130nm HSUPA RF, 180nm Analogue Baseband
Early-2009: 65nm Baseband, 65nm HSPA+ RF
Mid-2010: 40nm Baseband, 65nm LTE RF
Late-2011: 40nm Baseband, 40nm LTE/CDMA/WiMax RF
I don't know about you, but I'd gladly invest some money in that kind of R&D pipeline too :)
Judging by their rather brusque public comments about Qualcomm, I think they feel they have no need for any assistance from San Diego.
Your 65nm roadmap is spot on, so far as I know.
The 40nm, i agree, would be spot on, if achieved. And Knowles' chips always work first time!