Wireless Industry Analyst

• The 3G version, as I have already mentioned, suffers from a costlier chip also. The average price per 3G chip is closer to $25 as against single digit numbers for the Infineon chip in the iPhone today. Besides, the royalty structure for 3G amounts to about 28% of the average selling price of the 3G phone. Apple typically keeps its cost constant for more features. This will imply reduced margins from iPhone V2
• The power consumption issue does open up the question of who the baseband supplier will be. Will it be QualComm (QCOM), which has made strides in power management since iPhone V1? Will it be Infineon that has been confirmed to be developing UMTS (3G) baseband chipsets for atleast 2 companies? And if so, where does the recent licensing deal between Apple and Interdigital (IDCC) fit in?
• Who will supply the A-GPS capability for the iPhone?
  
• Would iPhone V2 be HSDPA-enabled?
  

• Feature integration
• Power management
• Ease of integration

• have better power management capabilities
• have smaller space requirements for the chip-set (which would also be enhanced by migration to the newer 65 nm technology) implying more room for music storage
• be backward-compatible with GSM networks. This is more a requirement for Apple given that most 3G networks are not pervasive and users are likely to be serviced by both GSM and WCDMA based on their location

• The device promises direct music transfer from the internet to the mobile, a no-no with iPhone.
• It is likely to have 3G versions during release. Whether it promises higher data-rates or not, at least there is a perceptual advantage with the 3G tag.
• Nokia has 1/3rd of the hand-set market share in the world. So even in places where the carrier lines up phones (eg. U.S.A as against in India), they should have negotiating power to leverage unprecedented deals with respect to retaining their own services etc.
• Nokia certainly has the brand-name to compete with Apple. The snob-value associated with an iPhone may not be there, but people will definitely trust a Nokia device enough to buy it.
• Nokia phones are not likely to be locked to carriers. This means more customers all around the world. They are gunning for instant pervasiveness as against the exclusiveness tag from Apple. They will have to price it below the iPhone or at least competitively to gain with this strategy.
• Though it is a brazen copy, Nokia is likely to learn from iPhone mistakes and present a better product. They will directly be competing with iPhone V2 when that comes up.

• QualComm's legal battles with Nokia (NOK) and BroadCom (BRCM) and the impending ITC ban: Based on the final rulings in much of these cases, QualComm chips may not be allowed to enter the US in the future and this will include an iPhone if it features such disputed chips. Besides, the royalty that it may have to pay the other parties may cut down the margins substantially on a per-chip basis. This is true in general, but a bad publicity with a highly visible product is the last thing the company wants.
• WCDMA royalties: If a 3G iPhone is available in 2008, about 2.5 million units will be sold that year. This number is expected to climb to about 17.5 million units in 2010. With a conservative average selling price of $350 and a royalty percentage of 6% (especially if Infineon does not have any underlying license deal), QualComm stands to make close to $50 million next year and as high as $340 million in 2010 just on iPhone sales. Compare this against $@ million a quarter, 7-year deal between Interdigital and Apple. This number appears exagerrated, but unless Infineon is able to work around the IP issue, the numbers stand.
  

• The $2 million increase in the guidance that seemingly is the basis behind the $56 million calculation may after all be royalty for the 2G phones sold through this quarter.
  
• On the 3G front, the royalty from IP itself may be higher, so that number may increase next quarter and beyond.
• Most importantly, in the event of Infineon getting the design win, iPhone will be using substantial contributions of Interdigital from both the software protocol stack and the design. Besides, Infineon has a per-chip royalty agreement for these two pieces.

The fourth part of my Interdigital series on Sramana Mitra's site.... Some of it is a re-cap of what I had written over three months ago here.

I concluded the last article in this series postulating that Interdigital will stand to gain not only from the licensing of essential patents to Apple but also from individual iPhone sales. To reason this out, let us take a quick look at the Infineon chipset in the current iPhone design-

The platform used was Infineon’s MP-EU. The platform itself supports UMTS (3G WCDMA FDD).

• Contrary to a myth, the current iPhone does not support 3G that can be enabled by a software switch. The chip used, SGOLD2-8876 is only capable of EDGE (a 2G enhancement). So, despite the platform’s capability, the iPhone is limited by the chips in it.
• MP-EU has Interdigital 3G protocol stack and baseband design and the King of Prussia-based Interdigital stands to get royalty for every 3G phone using the platform.

Based on this information, if Apple continues to use MP-EU for its next generation, upgrading its baseband chip alone to support UMTS, then Interdigital will immediately start to get per unit royalty. Of course, MP-EU does not support data-centric HSDPA which will give a better mobile internet experience. Infineon’s next generation chipset MP-EH supports HSDPA also and is covered under the 2006 expanded agreement with Interdigital. I think that MP-EH, along with S-GOLD3H, the HSDPA baseband solution that Infineon sampled last February is a good iPhone V2 candidate provided performance and interoperability issues are ironed out. There are at least two reasons that come to my mind when we seek to speculate on the heart of the next iPhone -

• Apple already has the MP-EU in the market and is testing it for stability and optimizing it for performance. The Marvell Wi-Fi solution had performance issues that were quickly ironed out. Similarly, Apple is conscious of the power consumption of the Wi-Fi chip and its effect on the total battery-life. I am guessing that the iPhone creators are already working to address such issues in the current platform. So, they are unlikely to prefer a new solution, at least in the short-term. In fact, in the absence of a stable HSDPA solution from Infineon, Apple may still opt to come out with just a WCDMA/UMTS version. This will require minimal changes to the current platform.
• The total cost of MP-EU and the 2 IFX chips in the iPhone is speculated at close to $16. Infineon can potentially offer Apple its 3G solution at a comparable cost (perhaps ~$20).

In summary, I speculate that Infineon will continue to be in the iPhone and therefore Interdigital also stands to gain substantially. In the sequel, I will take a quick look at my interpretation of Apple’s deal with Interdigital.

• The platform used was Infineon's MP-EU. This platform supports UMTS (3G WCDMA FDD).
  
• Contrary to a myth, the iPhone does not support 3G that can be enabled by a software switch. The chip used, SGOLD2-8876 is only capable of EDGE (a 2G enhancement). So, despite the platform's capability, the iPhone is limited by the chips in it.
• MP-EU has Interdigital 3G protocol stack and baseband design and the King of Prussia-based Interdigital stands to get royalty for every 3G phone using the platform.
  
• The 2G protocol stack is from Infineon itself but there is potentially some Interdigital IP involved in the chip+platform even without 3G. The license is retro-active to the day one of iPhone implying 2G royalty.
  
• The total cost of MP-EU and the 2 IFX chips in iPhone is speculated at close to $16.

As analysts and technology enthusiasts rush to dissect the iPhone to unravel the 'closely-guarded' secrets of the chips inside, it does not come as a surprise to me. To most wireless insiders, putting two-and-two together should have been easy. In most of my blogs, I have been taking for granted that the Infineon GSM chip is inside. The other component chips were also fairly guessable based on public stance of some of these companies and general word-of-mouth.

So, though it comes as a confirmation to those dis-believers, its a fairly redundant one at that. To me what is interesting is the iPhone V2 going forward. That is where the guessing game should be. So, all the hoopla about shares of these chip-makers going up are ratifications of the truth, as I have pointed out in my previous articles, I would not be bullish about most of these companies based on what they have in there today.

For the uninitiated, Infineon's GSM chip, Marvell's Wi-Fi, Samsung's memory and BroadCom's controller chip are in the iPhone. Intel, TI and National also had a share of the pie.

The alliance with Infineon has been a rather crucial one for Interdigital. The companies have been working together on 3G technologies in a strategic relationship since 2001. Last year, there were two important pacts that strengthened this alliance –

• Infineon licensed its GSM/ GPRS/EDGE S-GOLD 3 baseband modem and protocol stack software from subsidiary Comneon to InterDigital. This essentially allowed Interdigital to produce its own chipsets with 2G and 3G baseband modems. The agreement gives Interdigital the right to use Infineon’s 2G technology in its own modem offering or to sublicense the technology to third parties developing their own 2G/3G modem offerings. This deal facilitated the company’s ASIC plans in the 3G space. Interdigital successfully conducted initial evaluation tests on this combined 2G/3G modem with a state-of-the-art HSDPA advanced receiver solution this month. While Interdigital continues its march towards the smartphone market, Infineon will benefit from the licensing of components of the modem, the complete design, and of course, the sale of ASICs.
• In January, the two companies expanded their relationship to incorporate HSDPA development as well. Under this expanded agreement, the companies will continue the joint development of the 3G protocol stack software technology for Infineon’s 3G platforms. This agreement expanded the collaboration to include the data-centric HSDPA as well. This agreement also “amends the per-unit royalty rates to be paid to InterDigital for the sale of Infineon’s ASICs containing jointly-developed protocol stack software.”

While the first agreement gives Interdigital a foothold in the highly competitive ASIC world, the second agreement is also fundamental for its future outlook. Here is why –

• Interdigital gets a share for every Infineon 3G chip sold world-wide
• Infineon’s 2G chipsets are in today’s phones, most notably the iPhone

So, if iPhone V2 which will arrive in 2008 has the Infineon chipset, then Interdigital will stand to make money from every iPhone sold world-wide. In the sequel, I will look at why this is likely to be the case.

Will Qualcomm acquire Interdigital? This was the question I raised in my last article. I will try to take a quick look at the feasibility of such an event in this part of my series.

Interdigital has a market cap of about $1 billion today. The stock closed on Thursday, January 10 at $21.47, which is relatively low when compared to its 52-week high of $35.74. If my iPhone speculations are true, I anticipate that the company’s stock will climb back to these numbers when the iPhone V2 is out later this year. The strengths of this number include its per-unit royalty from its Infineon alliance and the ASIC business. The risks include the sustenance of the ASIC business and the litigations. I also think that if this deal were a reality, Qualcomm may have to cough up between $35 and $45 a share amounting to a deal worth $1.75-$2.25 billion.

The numbers certainly look staggering, but perhaps not for Qualcomm. The company closed on January 10^th at $37.99 with a market cap of about $62 billion. Interestingly as of Sept 30^th 2007, Qualcomm had cash and equivalents of $2.4 billion and short-term investments of another $4.17 billion. So, the San Diego-based company can certainly afford a big acquisition. The question is why should it acquire Interdigital and will it do so?

Why should it Acquire? Let me try and bring out the financial impact to Qualcomm here. Interdigital traditionally licenses its IP for a flat-fee. On the other hand, Qualcomm charges per-unit handset royalty as a percentage of its average selling price (ASP). The exact proportion of the ASP that Qualcomm will receive is the center its ongoing battle with Nokia. If Qualcomm acquires Interdigital before any such deal is finalized, it will get an upper hand in the discussions. A 1% higher licensing rate from Nokia is worth about $2.4 billion to Qualcomm. This figure, even when discounted at 8%, is worth about $1.9 billion. 1% may be on the higher end of a gain from IDCC’s IP portfolio to Qualcomm. But I have used these numbers to illustrate the higher worth these patents could be to the licensing model of Qualcomm than to IDCC itself. If Qualcomm is certain that the acquisition can guarantee such ballpark numbers from Nokia alone, the decision then will almost be a no-brainer.

Will it acquire? I am sure it will if it needs to. For one, Qualcomm has, in the past, never hesitated to make big acquisitions. Its 2000 acquisition of SnapTrack for its GPS capabilities is a good example. The company paid $1 billion in stocks to the San Jose-based SnapTrack for close to 50 essential GPS patents and its engineering team. Similarly, it paid a comparable amount to acquire Flarion in 2005 for the OFDMA patents and the team. While Interdigital does not bring in new technological capabilities like the other two did, the King of Prussia-based company certainly has the complementary IP portfolio that Qualcomm would love to have. Besides, with its recent spree of acquisitions, Qualcomm has sent a clear signal that it wants to expand and strengthen its leadership position. So money will certainly not be an issue if Qualcomm determines that Interdigital adds value to its profile.

It is true that my excitement about Interdigital is very speculative, especially with respect to its iPhone involvement. But that will only affect its valuation and hence the price for an acquisition. However, I am certainly confident that the synergies from this acquisition, if it happens, can cement Qualcomm’s future while stabilizing its current position.

I came across an interesting article at

http://wirelessdreams.wordpress.com

where the author mentions that the future technologies may not live up to the promise of high-speed and that better user experience with iPhone V2 and 3G is only a myth. While I partially agree with arguments and that 3G won't be the panacea for all the user woes, there is some reason for the expectations.

My argument is two-folds. One is that while the GSM networks around are fairly well populated and leave little room for data speeds, the HS and other 3G networks currently rolled out are nowhere near capacity. In other words, as long as these networks are not dense, there is a possibility that the initial users (I can't give a time-frame though) are likely to get an enhanced user-experience with respect to speeds. I am implying that the infrastructure in place for 3G far exceeds the current usage and hitting the density of the GSM networks may take a while.

Secondly, the argument with WiMax, I suppose, is cheaper infrastructure costs. This means that more sectors/base-stations can be laid for the same cost thereby supporting more users/higher data-speeds for the same investment. This then can also mean higher data-speeds coming from the future generations.

This being said, I will definitely buy the author's argument when the sectors become capacity-limited or in other words when there are more users in the network than what it was designed for.