Wireless Industry Analyst

Apart from the cellular ICs, Infineon has a low, non-negligible market share in the Bluetooth market. In collaboration with Global Locate (now Broadcom), Infineon also developed the Hammerhead II A-GPS solution for the cellular market. It also has a presence in the digital cordless and RF power markets.

While Infineon specializes in low-power chipsets and is looking to increase its wireless footprint across Europe and elsewhere, there are holes in its product portfolio. This is especially critical in the context of the global movement towards convergence devices. Firstly, the company does not have an application processor family that complements its rich and highly competitive baseband solutions. Secondly, it does not have a complete set of mobile connectivity products. With Global Locate getting acquired by Broadcom, Infineon’s future in GPS/A-GPS is unclear. Unlike the other leading cellular IC vendors, Infineon also does not have embedded WLAN solutions that will complete its current connectivity portfolio.

Infineon’s strategy seems to be to provide low-cost cellular platforms whose gaps can be filled by other vendors. The 2G iPhone platform is a good example. While its cellular ICs were from Infineon, the application processor was from Samsung, WLAN from Marvell and Bluetooth from CSR. This model will work fine as long as handset vendors are interested in picking the best-in-class components based on cost and performance. If, on the other hand, the trend drifts towards highly integrated single vendor solutions, the handicaps in Infineon’s mobile portfolio will result in the company losing its market share, potentially including iPhone designs 2009 and beyond. Infineon will not be in a position to compete with the likes of Qualcomm, Broadcom and STM, all of whom have a complete portfolio to build single-stop cellular platforms.

Given these loopholes in its mobile and wireless portfolio, mergers and acquisitions become viable paths for Infineon to stay competitive. In the sequel, we will discuss potential acquisition areas and targets before looking at its valuation.

The main products of this segment include power semiconductors, sensors and microcontrollers, silicon discretes, chip card and security ICs. Infineon’s ICs are used in power trains, body and convenience features such as air conditioning in your car, its safety features such as ABS, airbags and stability control, and its infotainment aspects. Industrial applications include power management and supplies, power generation and distribution, and industrial control. Customers include Avnet, Bosch and Siemens.

Market research firm Semicast reported that Infineon is the leader in the industrial segment ahead of STM and Renesas. Semicast estimates that Infineon has about 7.5% of the $20 billion industrial market. Further, this market is expected to grow to $33 billion in 2013 presenting Infineon with a good growth driver. Strategy Analytics ranks Infineon as the second largest supplier of ICs for automotive applications worldwide with about 9% market share. Additionally, Infineon is number one in the power semiconductors market with 6% share in 2006 according to an IMS research report.

Driven by the strong demand for energy efficient solutions, the industrial and automotive business will stay healthy. The company expects healthy growth rate in the industrial segment primarily based on its leadership position in power applications. It also hopes for roughly stable sales within its automotive business based on its ability to provide complete high quality semiconductor solutions integrating power, analog, mixed signal IC and sensor technology.

Like Semicast’s Colin Barnden notes, the challenge will be to hold on to the leadership position in the face of increased competition. The uncertain economy will make the stable industrial and automotive segment very attractive for semiconductor companies. The segment thus presents a good opportunity for Infineon subject to competitive risks. For the moment, I am willing to give Infineon the benefit of doubt that it will at least continue to maintain its market share here if not grow it modestly.

The picture below, taken from Infineon’s fiscal 2007 annual report, illustrates the root cause of Qimonda’s misery.

The DRAM prices dropped a whopping 29% in fiscal 2007 led by seasonal demand weakness, inventory build-up prior to the Windows Vista launch and capacity conversion from NAND to DRAM by the competition. The 44% increase in shipments and the continuing market diversification strategy were not enough to stave off these challenges leading to a net sales decrease of 207 million Euros in 2007.

Despite these issues, the DRAM business requires Qimonda to invest heavily in its research and manufacturing capabilities. In 2007, Qimonda announced plans to build new manufacturing facilities in Singapore and Malaysia. It also announced an agreement with SanDisk to jointly develop and manufacture Multi-chip Packages. Additionally, the company hopes to leverage its product and market diversification strategy to increase DRAM average selling prices (ASPs), profitability and return on capital.

Last week, Qimonda and Elpida, both top-5 DRAM suppliers, signed final contracts for a strategic technology partnership on joint development of DRAMs. Primarily a move to gain scale, this agreement encompasses a broad cross licensing of Intellectual Property that gives both companies higher design freedom while aligning their development interests. Elpida’s CEO is also quoted as saying that he was open to a share swap with Infineon.

Infineon wants to get out of the memory business. It announced recently that it wants to reduce its current 78% stake in Qimonda to below 50% by the 2009 Annual General Meeting. While not many will pay cash for this stake, share swaps with companies like Elpida may help constrict the supply and stem the free-fall in DRAM prices. With private equity unlikely to pick up Qimonda due to its unattractive cash flow, such share swaps may be the unfortunate reality that Infineon has to face today.

In summary, Infineon is unlikely to relieve itself of the memory business overnight. The huge write-off in the last quarter demonstrates that the company has realized this hard fact. It will have to swallow more losses despite its valuable R&D and manufacturing capabilities. From a valuation perspective, Qimonda will continue to have a considerable negative impact on Infineon’s share price.

For the fiscal year ending September 30, 2007, Infineon reported net revenues of 7682 million Euros at a gross margin of 20.7% and operating loss of about 3.5%. This is a decrease of 3% from the 2006 revenues of 7929 million Euros. The gross margin for 2006 was 26.2% while the corresponding operating loss was about 3.4%. Much like its European competitor STM, Infineon is also struggling well behind the industry average gross margin of 50.5% and operating margin of 18.6%.

The automotive and industrial business segment contributed 39% of the revenue while the communications segment accounted for about 14%. Qimonda, the struggling subsidiary producing memory chips, contributed close to 50% of Infineon’s 2007 revenues. Almost one third of Infineon’s products are consumed within Europe while Asia-Pacific accounts for another third.

Much of Infineon’s woes are tied to the struggling Qimonda. The commoditization of memory chips has caused their prices to fall considerably. While the cost of owning and operating memory fabs is high, the RoI is just not there due to unfavorable market conditions. Severely impaired in its ability to generate shareholder value, Infineon recently declared its interest in Qimonda as ‘assets held for sale.’ To give more clarity on its core business, the company will now report revenues only from its continuing operations. The potential gains or losses due to the change in the fair value of Qimonda will be reported as discontinued operations. In accordance with this practice, the company wrote off 1.3 billion Euros last quarter.

In the recent quarter that ended March 31, 2008, the company reported net revenues of 1049 million Euros from its core segments. This is an increase of 7% from the corresponding quarter in 2007. The gross margin for the quarter was 35% while the operating margin was about 4.5%.The Qimonda write-off in the balance sheet demonstrates the quandary that Infineon is in. It wants to dilute its stake in Qimonda and get out of the memory business. But apparently, the loss-making subsidiary does not have many takers. Infineon’s handling of Qimonda in the next few years will be an important component of any valuation analysis. It will hence be illustrative to take a look at this relationship in the sequel.

Infineon shot to the spotlight in the wireless world as it was discovered to be at the heart of the sensational first generation iPhone. I also think that Infineon will continue to hold this socket in the 3G iPhone to be released on July 11^th. I am betting that Infineon’s XMM6080 platform and its PMB8878 chip (formerly known as S-GOLD3H and now marketed as X-GOLD608) will be in Apple’s darling product that was announced amidst much fanfare and expectations recently.

While the iPhone has been Infineon’s public face around the world for the last year or so, its business interests are much broader spanning automotives, industrial applications, and memory among others. The company was founded in the summer of 1999 as a wholly owned spinoff of Siemens AG following the latter’s huge losses in the fiscal year 1998. Infineon Technologies AG was listed as an IPO on the Frankfurt and New York Stock Exchanges in 2000. Today, iSuppli ranks Infineon 10^th in the list of worldwide semiconductor suppliers.

The company is broadly divided into three business segments –

Automotive, Industrial and multimarket segment designs, develops and manufactures and markets semiconductors and complete system solutions for automotive, industrial, security applications and customer-specific applications. The continuing demand for high power products for industrial applications, increased sales of energy efficient products and an increased demand for government ID applications are attributed as growth drivers for this segment.

The Communication segment offers a range of ICs, other than semiconductors and complete system solutions for wireline and wireless communication applications. Wireless chipsets are a mainstay of this unit. Beaten up over the past two years by the insolvency of BenQ’s German subsidiary, the company is trying to recoup by building its IP and product base mainly through acquisitions. Besides Apple, Infineon phone platform customers include ZTE, Panasonic and LG.

Qimonda was legally separated from Infineon in May 1, 2006 into a stand-alone company manufacturing memory products. This segment has been Infineon’s Achilles’ heel from the day one. In fact, Infineon’s inception was forced because memory price erosion led to a $674 million pre-tax loss posted by Siemens Semiconductor in 1998. Even in 2007, a drastic 29% decrease in DRAM prices hurt Infineon’s prospects. The German chipmaker, realizing this bottleneck is starting to dilute its ownership in Qimonda.

With this brief overview of the company, we will look at the company’s financials, business segments, strategy and valuation in the sequels.

Firstly, IDCC will get licensing fees from Apple as per agreements signed last year. The agreement covers 2G and 3G iPhones to be sold in the near future. So, irrespective of whose 3G components are in the iPhone, IDCC gets money from Apple.

Secondly, there will be money coming if Infineon’s 3G baseband is in the iPhone. In light of the recent product announcements by Infineon, I need to be careful when I make that statement. So, let me step back and reflect on the impending 3G iPhone. I wrote in the past that I expect to see the MP-EH platform with the S-GOLD3H baseband chip in the 3G iPhone. The company has rechristened its products since. The HSDPA capable MP-EH platform is now the XMM 6080 while its 3G baseband chip - PMB8878 - previously known as the S-GOLD-3H (H indicating HSDPA capability) is now the X-GOLD 608. Read more on this subject here and here.

Further, the 3G software stack that will be used in the MP-EH/S-GOLD3H combination was jointly developed by InterDigital and Infineon’s subsidiary Comneon. So, IDCC will get per-unit royalty from Infineon for use of the stack. While the software collaboration is public, I am inclined to speculate based on recent modem performance data that IFX 3G chips use some baseband receiver design IP from IDCC. This can be an additional source of per-unit royalty for IDCC.

Recently, Infineon announced an array of X-GOLD61x HSDPA/HSUPA products that drastically reduce on power and space apart from catering to multiple phone market segments. Furthermore, the company also announced specifically that the stack for the X-GOLD61x solutions is internally developed. While this announcement gives more design control to Infineon, it implies that IDCC will not receive per-unit royalties for the software stack on the 61x series. On the other hand, I am still bullish (read speculation) about the use of IDCC’s physical layer receiver IP in all of Infineon’s 3G offerings. Also, while I think these new products are part of Infineon's push to hold on to Apple for future designs in the face of increasing competition, I would be greatly surprised if I see any of these 61x chips in the first 3G iPhone.

In summary, look to see if the iPhone teardown reveals Infineon’s PMB8878/X-GOLD 608/S-GOLD-3H baseband. So, the best case scenario for IDCC is for the iPhone to have a heart of Gold! Even otherwise, the King of Prussia-based company will gain from Apple’s licensing fees.

[Long IDCC at the time of writing]

The IDCC-NXP connection: NXP has licensed IDCC’s 3G HSDPA ASIC design for its PNX6712 chip. The companies entered into this agreement for integration into NXP’s Nexperia cellular chipsets as early as August 2005. Talking about this during InterDigital’s 1^st quarter 2008 conference call, its CFO Scott McQuilkin said, “We also completed our delivery of HSDPA technology to NXP and ASIC is now moving into production.”

The NXP-STM connection: STM and NXP entered into a JV last month merging their wireless businesses to acquire scale. You can read my extensive coverage of this event here and here. The formation of a solid number three player stints IDCC’s abilities to sell its SlimChip solution directly. But in the bigger scheme of things, it spells good news for the King of Prussia-based company as the JV widens its footprint. In the most recent quarter conference call, IDCC’s CEO William Merritt resonated this point saying, “Frankly we looked at the combination of ST Micro and NXP and said, ‘Great, they’re getting stronger and they have our IP’ and we love our licensees to be strong and pushing hard into the market..”

The STM-Nokia connection: Nokia last year decided on a multi-vendor sourcing strategy. It now sources 3G from both TI and STM. Additionally, it also transferred about 200 chipset design engineers to STM. This not only signals Nokia’s move away from chipset design, but also presents STM with the opportunity to grab more 3G business at Nokia. It is devoting R&D resources to baseband design and NXP’s team and product are an active part of the equation.

The performance conundrum: Data communication necessitates a better receiver design. Despite the design resources from Nokia, STM will take time to build a competitive solution on its own. On the other hand, the design that NXP has licensed from IDCC is one of the top-performing receivers today. Nokia cannot afford to lag in 3G performance especially considering its high-end mobile market competitors have good solutions. Samsung sources from Qualcomm and Infineon while Apple is expected to have an Infineon chipset in its impending 3G iPhone. So, for the moment, the IDCC design seems to be the most accessible ‘best’ 3G solution for Nokia.

In summary, as STM continues its 3G ramp up, it may use NXP’s design in its platform in the short-term. IDCC, by virtue of its license agreement with NXP, will directly benefit from this move. What is more, this will give IDCC a chance to be part of Nokia phones. Now, that begs another question. Are IDCC and Nokia factoring this in their licensing discussions?

Firstly, Sehat Sutardja, Marvell’s CEO, had this to say in the company’s F1Q09 conference call –

“During the first quarter we achieved what I believe to be a very important milestone as we began volume shipments of our HSDPA communication processor to a key smartphone customer. We expect a steady ramp to high volume production throughout the remainder of the year.”

With the name of this ‘key smartphone customer’ not revealed, Apple and iPhone immediately got tagged to this remark. To me, the matter-of-fact statement appeared to reflect Marvell’s continuing relationship with RIMM.

The latest Blackberry Bold 9000 features Marvell’s Tavor platform. Tavor is a single-chip solution that combines a HSDPA baseband processor with a 624 MHz applications processor. The timelines of the Bold launch and Marvell’s shipment dates seem to corroborate as well. The RIMM angle, coupled with the difficulty in launching and testing a new platform with Marvell for the existing form-factor iPhone makes me believe that the Santa Clara-based company will not displace Infineon in Apple’s darling phone.

I will, however, not dismiss the possibility of Apple launching a second, smaller form-factor 3G phone with Marvell’s solution. Marvell’s Tavor may be ideal for a low-cost phone from Apple. The single chip will eliminate the need for a separate application processor. It will save space and power. Hence, it (or any single chip solution for that matter) will be a preferred solution a low-cost iPhone, if there is one. I also anticipate that Marvell will be able to bundle Tavor with its WiFi solution giving it a price advantage. So, in the eventuality of two iPhone models being launched, it is possible that both Infineon and Marvell have design wins at Apple.

This brings me to the other iPhone related news – the Infineon warning. The German company recently warned that it has received lower than anticipated orders for a project to supply HSDPA chips. Going with the premise that Infineon is indeed the 3G supplier for the iPhone, it is hard to tell if this warning pertains to Samsung or Apple or another customer. If it is for Apple, then, contrary to what is being written elsewhere, I don’t think it signifies any major delay in the anticipated launch of the 3G iPhone.

This may, however, signal one of two things. The first possibility is that Apple may not market the 3G iPhone as widely as anticipated (at least initially.) The second possibility, which ties up with the Apple angle to the Marvell statements is product mix. Apple may be planning on a staggered, complementary launch of two 3G iPhone models. This, in turn, could be the reason behind the lower volumes shipped by Infineon and the perceived secrecy associated with Marvell’s ‘key smartphone customer.’

In summary, I think Infineon will be in the 3G iPhone. While I feel that Marvell’s statements pertain to RIMM, I also see the company as a very good candidate for a second iPhone design that may also launch soon.