Qualcomm (QCOM) is the world's leading smartphone chip vendor. It is simultaneously the largest discrete modem vendor and the leading vendor of applications processors, holding over 50% market share. Since Qualcomm is a fabless semiconductor company, it relies on third parties -- principally, Taiwan Semiconductor (TSM) -- to build the chips that it designs. A recent report from the Business Korea suggests thatSamsung (SSNLF) is trying hard to get Qualcomm's foundry business, but is Qualcomm falling into a trap?
Samsung isn't just a foundry -- it's a chip designer and device vendor
Unlike TSMC, Samsung has chip design teams that compete directly with Qualcomm's. Indeed, Samsung develops the following in-house:
- Cellular baseband
- RF Transceiver
- Smartphone/tablet applications processors
- Processors for Smart TVs
- Server processors
Not only does Samsung develop all of these products in-house, but it actually is the leading vendor of the consumer products that these chips go into. It's the world's highest volume smartphone vendor, second-highest volume tablet player, and the world's largest TV vendor. That said, Qualcomm has been able to win a significant number of sockets in Samsung devices with its processors, including some variants the flagship Galaxy S5 and Galaxy Note products.
Samsung is competing with Qualcomm
There's a fundamental problem with Qualcomm enabling Samsung to be a top-tier foundry. Given how R&D-intensive these semiconductor manufacturing processes are and given how expensive it is to build a next-generation factory, there will only be a select few companies that can successfully/profitably build leading-edge silicon. Taiwan Semiconductor is one of them and Intel (NASDAQ: INTC ) is another, but beyond that is an open question.
Intel is unlikely to really ever become a general purpose foundry, so if Samsung were to ultimately end up being the other major foundry, it would have pretty significant control over much of the fabless semiconductor ecosystem. The problem here is that as Samsung's foundry business grows, Samsung can more freely invest in the development of its own silicon for its own devices, which have majority market share. At some point, Samsung could just keep using its own silicon inside of its products, competing with its very own foundry customers for sockets that Samsung controls!
Qualcomm would be making a mistake enabling Samsung
While Qualcomm's bread-and-butter business is the set of royalty checks it collects from anybody who sells a 3G/4G device, its semiconductor ambitions are strong. If Qualcomm moves a significant portion of its leading-edge business to Samsung, it may be able to get aggressive pricing in the near term, but long term, Samsung has been very clear that it plans to develop its own chips and get those chips into as many of its devices as possible.
This means that unless Qualcomm can stay several steps ahead of Samsung's product definition and design teams, it stands to ultimately lose many of the high-margin/high-end sockets that it currently owns in Samsung's products. If Samsung shows signs that it's willing to stand down on the development of its own chips (which it has given no indication of doing), then a heavier reliance on Samsung's foundry would appear to make more strategic sense.
Foolish bottom line
Qualcomm's execution in mobile has been nothing short of incredible, and has led to the company winning many key sockets, even within Samsung's devices. However, as long as Samsung continues to drive its own chip design efforts, and as long as it continues to be one of the largest vendors of handsets, a heavy reliance on Samsung's semiconductor foundry by Qualcomm would appear to be a strategic error on Qualcomm's part. In the short term, Qualcomm gets second sourcing and probably more aggressive pricing, but longer term, Samsung's chip teams are probably going to aggressively aim to cut Qualcomm out of every socket possible, replacing them with Samsung-designed parts.