Stay on Top of Enterprise Technology Trends
Get updates impacting your industry from our GigaOm Research Community
The Asian chip manufacturers are getting together with NTT DoCoMo, the Japanese cell phone operator to create a joint venture to build mobile phone chips. Samsung, Panasonic, Fujitsu Limited and Fujitsu Semiconductor have said they will combine to create a new chip technology for mobile communications. From the press release announcing the joint venture:
The joint venture company, leveraging the six investing companies’ strong backgrounds in cellular communication technology and vast experience in application specific integrated circuits (ASIC) design and foundry manufacturing, will develop feature-rich, small-size, low-power-consumption semiconductor products equipped with modem functionality. The joint venture company will focus on developing products for LTE and LTE-Advanced mobile communication standards. The products will be sold in markets globally.
This means these companies are confident they can build a better chip than industry giants such as Qualcomm (s qcom), but it’s also an admission they must work together if they don’t want to miss the wireless opportunity. That’s an opportunity chip firms can’t afford to ignore.
These companies will establish the joint venture by the end of March 2012. It looks like they will integrate the application processors with the modem and other silicon on a system on a chip. Qualcomm is well-known for its mobile chip integration capabilities. Beyond Qualcomm, the announcement is a reminder of how mobile is redefining everything, including the staid world of chips, where former mortal enemies are now part of a joint venture.
The consortium is also a tactical admission by the Japanese that they alone don’t have the technical wherewithal to face the mobile onslaught being unleashed by U.S.-based companies such as Broadcom, (s brcm) Nvidia, (s nvda) Qualcomm, Google (s goog) and Apple.(s aapl) Japanese handset makers are increasingly marginalized in the global market because of their own unique mobile ecosystem, which prevents them from achieving large economies of scale.
Mobile is not PC
When PCs and servers ruled the roost, Intel (s intc) has a lock on the market thanks to its x86 architecture. It let AMD (s amd) follow, but never let it lead, creating a dynamic for the industry that has led to a lack of innovation in hardware beyond the needs of the mass market. AMD (S amd) pushed the envelope with innovations such as HyperTransport, and being the first to build 64-bit chips, but it was never able to outspend or out-compete Intel.
The mobile world is different. Chip firms start with the same basic ARM (s armh) instruction set licensed by the British company of the same name, and then innovate from there. There are hundreds of ARM licensees in various segments of the market. Some companies integrate their application processors (the brains) with the radio (the voice) of the phone, while others use separate chips. Qualcomm, for example, is famous for integrating its Snapdragon processor with its radios, and does it well enough that it has become the top baseband chip vendor in the world. But this is a game anyone can play, because no one has locked down the license for the architecture like Intel had with x86.
Even Intel is seeking a baseband advantage, buying up Infineon’s wireless business in January so it had the radio component to sit next to its Atom chips. The jury is out on Intel making it in this competitive mobile world, but it can’t afford not to try. Meanwhile, formerly huge players, such as Texas Instruments, (s txn) which exited the baseband business in 2008, seems to be losing the marketing battle with its OMAP line of chips. And players such as Nvidia (which bought its own mobile radio company) and Broadcom are seeking their entrée into the market as they realize the mobile device is going to be bigger than the PC ever was.
Back to DoCoMo & Co.
The news that NTT DoCoMo is eager to get in on the mobile game with some kind of joint venture is odd only in that an operator is leading the pack. However, Apple’s success with its A4 and A5 processors shows how deep vertical integration (owning everything from the chip to the app store) can profit a company, and Apple may also indicate why NTT DoCoMo is involved. Currently, Apple’s A5 chips are manufactured by Samsung, although Samsung’s rising prominence in the Android ecosystem has frayed the relationship. Perhaps NTT DoCoMo helps shield Samsung from the wrath of Apple as their co-opetition becomes more fierce.
And Samsung has long worked on radio chips to help it decrease its reliance on Qualcomm, so the idea that it’s part of this venture isn’t all that crazy. The new joint venture isn’t fully baked yet, but NTT DoCoMo said it will invest 450 million Japanese yen ($5.8 million USD) to establish a wholly owned subsidiary, called Communication Platform Planning Co., Ltd. headed by CEO Mitsunobu Komori. Komori is the EVP and CTO of DOCOMO. As the head of a chip platform and with the experience from being at mobile operator responsible for delivering some of the world’s most advanced handsets, he knows what consumers want, and clearly this joint venture is hoping it will deliver it.