High-stakes partnering: Who has the hole card?

It seems there have been a number of tech partnerships of late in which it's tough to figure out which partner is playing from a position of strength and which might be bluffing. Like poker players, tech companies play all-in for their own benefit. But partnerships often obscure the win-at-all-costs mentality. Take the Qualcomm partnership with Samsung announced in late November. Samsung, the handset maker, has been one of Qualcomm's biggest customers. Now, Samsung will be a fab partner, building cell-phone chips for Qualcomm. Is this a case of Qualcomm simply finding the best manufacturing partner out there? I'm not so sure.

Surely some amount of the success that Samsung has enjoyed in the handset business comes from the company's decision to endorse Qualcomm's CDMA technology. But CDMA-handset makers have paid a premium relative to their GSM (Global System for Mobile communications) counterparts because of the vast CDMA intellectual property that Qualcomm owns. Likewise, CDMA carriers have paid the Qualcomm tax as well, albeit they've arguably enjoyed more efficient networks.

Some in the CDMA camp—notably Samsung and carriers in Korea—seem determined not to be beholden to Qualcomm—or anyone else—for next-generation standards. As I covered in a recent feature story, Korea has actively sought to develop the country's own 4G mobile standard, WiBro, and WiBro may or may not align precisely with a WiMax standard (Reference 1). The Korean contingent was especially miffed when Qualcomm cut China a lower cost deal to win current-generation mobile business in what industry insiders commonly consider the biggest opportunity in the global market.

So, Qualcomm faced a partner that could've quickly turned into a competitor. Samsung has refined its IC-manufacturing capabilities using flash memory as a driver, and the company has become world-class in every way. Samsung could clearly be a threat in next-generation handset chips.

Now, however, Samsung will be making chips for Qualcomm. The companies haven't specified the scale of the deal. Qualcomm has long been among the biggest fabless companies, relying primarily on TSMC as a foundry. You could argue that Qualcomm simply needs an alternative source and additional capacity. In fact, that scenario is likely true. Presumably, Qualcomm will have additional opportunities as GSM carriers roll out wideband-CDMA networks and as some portion of Qualcomm's IP portfolio enters the larger GSM segment.

There are also many reasons to believe that Samsung is an excellent choice as a fab partner. The company has a solid technology base today. And Samsung is allied with IBM, Chartered, and Infineon to develop 65-nm- and finer-grain processes. Essentially, a relationship with Samsung could blossom into a manufacturing relationship with all four partners.

Still, I keep coming back to Samsung as a threat to Qualcomm and as a company unhappy with its share of CDMA cellular profits. By making a deal with Samsung, Qualcomm can keep the emerging giant solidly in the CDMA camp. Samsung will get a chance at a share of the profits as all flavors of CDMA proliferate. Qualcomm might even win over Samsung for support for Qualcomm's MediaFlo technology for video multicast to handsets despite the fact that Korea is already deploying competing technologies.

Then again, I don't think Samsung is all-in with this one deal. In fact, the company just demonstrated a WiBro-based phone and PDA at the 2005 APEC IT Exhibition (www.apecitkorea.org/eng/index.asp).

I hope you enjoy Bill Schweber's article on quantum cryptography. Bill has always produced excellent articles, and this one is no exception. And it's on a fascinating subject. Bill, however, has decided to leave EDN for a new challenge. We will miss him (and I'm sure you will, too), and we thank him for all of his years of great work.