Electronics Weekly Electronics Design & Components Tech News
Broadcom will “close the gap” with Qualcomm in the baseband market within a couple of years, Broadcom CEO Scott McGregor tells the Wall Street Journal.
This is an exceptionally ballsy aspiration for the CEO of a company which is not even in the top five baseband suppliers.
No.1 in the $3.6 billion baseband market is Qualcomm with about 42% share; No.2 is Intel on around 17% share and No.3 is MediaTek with 14% share followed by TI and ST-E.
So Broadcom has a hard row to hoe. On the other hand it was showing triple digit unit shipment growth in mid-year. As, indeed, was China’s Spreadtrum.
Helping Broadcom’s ambitions are the rise of LTE, the decline of CDMA and an internal technology development push which will enable it to deliver a smartphone BOM of $100 in 2012.
McGregor says: “$100 to $200 unsubsidised smartphones are very interesting to us.”
However fablessness rears its ugly head at this point. Eric Brandt, Broadcom CFO, acknowledges that: “TSMC is typically one process node behind Intel.”
Furthermore, even though Broadcom uses other foundries than TSMC, Brandt states: “Even on 40, it’s hard to find a supplier of 40 other than TSMC”.
TSMC, of course, is also Qualcomm’s foundry, and Qualcomm is often the lead process development partner on TSMC’s low power process which may give it priority rights in a tussle over capacity allocation.
So, in the fight for baseband market share, TSMC may have quite a big part to play in deciding the outcome.
Currently Broadcom’s strength is at the low-end but, as it moves to make highly-integrated, low-cost chip-sets for smartphones, and as the slowing of CDMA inevitably weakens Qualcomm’s grip on the baseband market, and if there’s plenty of 28nm capacity at TSMC so Broadcom is not starved of wafers, then Irvine could significantly diminish the gap with San Diego in 2012.
Broadcom says they shifted focus to smartphones this year. I am surprised they werent already doing this last year. The way Broadcom might be able to catch up to Qualcomm is the fact they have 22 product lines and provide an end to end internet solution. I am not an engineer but if all the various BC chips in the network, handheld, and home, including the 3g/4g phone-tablet, fiber, ethernet, backhaul, etc were to tightly integrate better then the various hodge podge solutions it might make for better bandwidth performance and thus
BC gains the upper hand. Tall Order i guess.
@Arun
great summary of BRCM’s phone capabilities, But, you still have the question of, in exactly which markets will they take share? Lets assume they mean units rather than revenue, then it is clear they must attack BRIC ( especially China and India) Yikes that means going toe to toe with Spreadtrum at the low end and the Taiwan twins for low end Smart-phone.
Otherwise they need to capitalize on NOK or whoever takes their share in Europe. maybe they buy TXN/STE and get it that way…
Five quick points:
1) BCM21654 looks pretty good for low-end smartphones, very similar specs to the ST-Ericsson U4500, both based on the Cortex-A9. However Qualcomm uniquely has an even lower-end and cheaper option with the Cortex-A5 in the MSM7227A on 45nm and that’s already shipping. They also have a very strong roadmap for the low-end with the 28nm MSM8225 which will have higher performance than even the Galaxy Nexus today.
2) Broadcom has an advantage in terms of better RF integration expertise. They are working on integrated 40nm cellular RF and I expect they’ll couple that with a dual-core Cortex-A7 which could be a very good combination and match or beat the MSM7227A’s Bill of Materials while delivering much better performance.
3) Broadcom acquired Beceem which used CEVA-XC for their LTE chip. It’s a very good SDR solution and should allow them to develop a 2G/3G/4G baseband that’s significantly smaller than Qualcomm. On the other hand, Broadcom’s multimedia might (or might not) be a weak point.
4) Broadcom will exclusively focus on High-K nodes (HPL/HPM/HP for TSMC) at 28nm while Qualcomm will focus primarily on SiON (LP for TSMC). And they obviously have huge volumes as well. I don’t expect capacity allocation to be a problem.
5) Qualcomm has always done better than many expected despite strong competition. Never, ever, count them out.
French pride? Didn’t the Germans make short work of that a few years back?
Would French pride allow that to happen , Another Anonimous?
‘However fablessness rears its ugly head at this point. Eric Brandt, Broadcom CFO, acknowledges that: “TSMC is typically one process node behind Intel.” ‘
Well I respect Scott a lot but his CFO needs to understand that Intel’s (i.e Infineon’s) baseband chips are made at … TSMC.
Anyway at least getting back into the top 5 should be easy given STE’s strategy 🙂
Is Broadcom going to buy ST-E ?
BRCM will needs lots of luck with that strategy. I’d say they must have a TDSCDMA product before they have a chance at unseating QCOM. Means acquiring the IP (plus lots of work) or buying an existing TD supplier….Hmmm, maybe I smell an investment opportunity….