Thursday, February 19, 2009

QCOM: technical vs. fundamentals

Normally I don’t write about Qualcomm’s stock, but Qualcomm’s stock value got about the most prominent discussion possible on Thursday.

On his nightly CNBC show, retail stock pundit Jim Cramer spent several minutes Thursday night analyzing the last few months of momentum on Qualcomm’s stock, in his weekly segment called “Sell Block.” A written summary is provided on the Mad Money website (and also in a separate article).

He started by talking about Qualcomm’s fundamental business, including a simplified version of its business providing IPR and chips for cellphones. Cramer said: ”It reminds me of the old Intel inside ads. It should be Qualcomm inside.“

But the impetus for the discussion was a very negative report from technical analyst Helene Meisler. To briefly summarize
  • Any rise in Qualcomm’s stock price is blocked by its January peak.
  • The upward trends that started earlier this year has stalled.
  • When there is higher stock volume, the trend on those high volume days is “telling the truth,” and Qualcomm had some ugly down high-volume days.
  • The stock is heading down towards $30 and won’t break $40 again. (The stock closed at $33.84 Thursday).
Cramer, on the other hand, argues that the business fundamentals favor QCOM. It has been taking share against its rivals. It earns $4-$8/phone royalties (average phone price $200), with 90% margin. Worldwide 3G penetration to go from 40% to 70-80% over next 3 years. This week, Qualcomm even won cooperation from former archenemy Nokia.

Cramer was also not surprised by the lower CY 2009 earnings guidance that Qualcomm provided with last month’s earnings release. To Cramer (and to me), there is the implication that Qualcomm assumes that the economic recovery is pushed back at least to 2010 — a realistic assumption.

Cramer thinks it’s a good deal because it’s down from its $56 peak (last August 14). As Cramer said, “A lot of the bad news is baked in” praising its “pristine balance sheet.”

I don’t know which one is right, but I know they both can’t be right. If over the next 6 (or 12) months, the stock never breaks $40 and ends below $30, clearly Meisler. If it ends above $40, Cramer is right. If it just muddles along between $30 and $40, Meisler isn’t wrong, but that’s a pretty weak sell call — particularly if (as I expect) the broader market is sideways or down for the next 18-24 months.

Note: Cramer owns QCOM for his charitable trust. I do not own any shares.

Tuesday, February 17, 2009

Qualcomm's other shoe

After 10 years in which Symbian never supported CDMA or any Qualcomm chipsets, why did Qualcomm join the Symbian Foundation last week? As the EE Times reports, the other shoe dropped this morning:
BARCELONA — Both ST-Ericsson and Qualcomm Inc. have revealed partnership programmes with Nokia based round reference platforms that will use the Symbian Foundation's software.

Qualcomm's deal focuses on developing UMTS mobile devices, initially for North America, that will be based on the S60 software on Symbian OS, running on the San Diego, California- based chip maker's latest MSM7xxx-series and MSM8xxx-series chipsets targeting wireless broadband.

The companies, for long arch rivals due to long standing patent infringement law suits that were finally settled last year, say the first mobile devices based on the collaboration would be expected to launch in mid-2010 and be compatible with the forthcoming Symbian Foundation platform.
Nokia currently has a very limited relationship with US carriers. “UMTS … for North America” means selling S60 smartphones either to AT&T or T-Mobile, competing for shelf space with the iPhone or gPhone respectively. Certainly, outside the US the Nokia smartphones are sold side-by-side with the iPhone by the same carriers.

EE Times plausibly argues that this cooperation is because Nokia last summer finally resolved its patent fight with Qualcomm. In November, analysts told EE Times they thought it unlikely that Nokia would go so far as to buy chips from Qualcomm.

It seems particularly odd that Nokia would start its partnership with Qualcomm with W-CDMA phones — given its sells hundreds of millions worldwide — but perhaps it is just trying to prime the pump. Presumably the next step is to use QCOM chips to sell more CDMA phones in the US, although so far Nokia is outsourcing that work to ODMs.

On a somewhat related note, Qualcomm CEO Paul Jacobs was quoted in Barcelona by CNBC as saying demand for smartphones remains strong. However, unlike his predecessor (Jacobs père), Jacobs fils is a perpetual optimist.

TI and InterDigital ally against Qualcomm

In connection with GSM Mobile World Congress, TI and InterDigital have made announcements reflecting an alliance between them. The new development is helping InterDigital make its transition from being an IP licensing company (that a few rivals have called a patent troll) to a firm that makes 3G chips to subsume its IP.

Not surprisingly, InterDigital is more enthusiastic about the alliance than TI is. (It’s fun to read these tea leaves — as someone who ran a little company that did joint press releases with big companies — the message is pretty clear). InterDigital issued a press release trumpeting its W-CDMA (i.e. HSPA) modem option and how it works with the the TI OMAP 3 and OMAP 4 processors:
"We are pleased to be the wireless modem supplier for TI's advanced OMAP 3 platform. Our high performance HSPA modem offers instant mobile broadband connectivity, accelerating the development of compelling new applications," stated Mark Lemmo, Executive Vice President, Business Development for InterDigital. "Available as a 3G modem option, InterDigital's SlimChip MID Module has been pre-integrated with the Zoom OMAP34x-II MDP, allowing mobile application developers and OEMs to fully exploit the rich capabilities of this platform."
The press release goes on to quote Bill Crean, Strategic Marketing Manager, as praising “ The flexible architecture of the OMAP 3 platform allows it to easily connect to InterDigital's SlimChip MID Module.”

Meanwhile, the TI press release allows InterDigital space with other vendors who praise themselves and the OMAP 3. The only mention of InterDigital comes from InterDigital itself:
"We are pleased to be a wireless modem supplier for TI's advanced Zoom II mobile development platform. Our high performance HSPA modem offers instant mobile broadband connectivity, accelerating the development of compelling new applications," stated Mark Lemmo, Executive Vice President, Business Development for InterDigital. "Available as a 3G modem option, InterDigital's SlimChip MID Module has been pre-integrated with the OMAP34x-II MDP, allowing mobile application developers to fully exploit the rich capabilities of this platform."
Eleven months ago, blogger Vijay Nagarajan predicted on his own blog (and later on Seeking Alpha) that an alliance with InterDigital would help TI:
TI is the market leader for 3G application processors with its OMAP product-line. But it does not currently have a standard 3G baseband solution. The company’s OMAP roadmap merely has placeholders for future merchant ICs with 3G baseband. Its mammoth market share comes from custom chips it develops for Nokia and Motorola among others. This position is, however, challenged by the multiple sourcing trend that the handset vendors are now adopting. Left behind in the 3G baseband race, not only by Qualcomm but also by Broadcom, Infineon, InterDigital etc, TI is finding itself losing mobile share. The aggressive strategy and the niche product positioning by the competition is not helping its cause either.
So although Qualcomm has a unique combination of technologies, its rivals are able to ally to produce a competing combination. (We call this “open innovation.”) I don’t know how this will impact Qualcomm’s profitability in the long term, but it seems as though the increased competition isn’t going to help.

Monday, February 16, 2009

Qualcomm joins Symbian Foundation

Cross-posted from OpenITStrategies

In advance of this week’s GSM Mobile World Congress, Qualcomm has joined the Symbian Foundation (as have HP and MySpace). Qualcomm of course is interested in supporting CDMA around the world, in addition to its dual mode 3G processors.

Rumors of Symbian CDMA support date to at least 2004. Nokia had once sponsored an adaptation of Symbian S60 to work with CDMA – presumably to gain access to the US market and segment where it’s had a relatively weak presence.

Back in 2005, MobileTracker reported that Nokia had won FCC approval for the Nokia 6638:
The FCC has approved (FCC ID QMNRM-18) the Nokia 6638, the first CDMA Series 60 phone. Visually, it looks just like the Nokia 6630 with the addition of a [very long] antenna. Since the 6638 is a CDMA handset, Nokia is most likely aiming for a US release.
The phone was compatible with Verizon’s two CMDA bands, 800 MHz and 1.9 GHz.But the phone apparently never made it to the market.

Qualcomm has already joined the Android Foundation. Apparently Qualcomm has worked on a Windows Mobile prototype that works on Qualcomm chips, and its chips power several BlackBerry models that are dual cdma2000/W-CDMA, including the Storm and the 8830 world phone.

Now Qualcomm has committed resources to the major smartphone operating systems — pretty much everything outside the iPhone. From now on, we’ll see how many operators and manufacturers use Qualcomm chips for their Android, Symbian or Windows Mobile smartphones.

Friday, February 13, 2009

Korea 3G finally shifts to W-CDMA

For 2G, Korea was exclusively cdmaOne, the only country in the world where this was true. Operators quickly got into 3G by offering cdma2000 on their existing frequencies, but the government licenses specified W-CDMA for new 3G frequencies that (IIRC) were compatible with the Japanese selections. As with anything else in Korea, each round of choices was designed to protect (or at least help) the export efforts of Korean companies.

The website Cellular News reported Thursday that in Q4 2008, the total W-CDMA subscribers in Korea finally passed those for cmda2000: 16.5 million vs. 14.9 million. For me, this marks the end of an era for Qualcomm’s original success story in Korea, marred only by the periodic royalty complaints.

Thursday, February 12, 2009

CDMA success brings Chinese royalty gripes

Taking a page from their Korean counterparts, in the face of rapidly growing sales, Chinese cellphone manufacturers are airing complaints through local propaganda organs in hopes of lowering their contractually obligated CDMA royalties.

Last year, a reorg of China’s telecommunications industry brought dominant landline provider China Telecom the opportunity to offer 3G mobile services using Qualcomm’s cdma2000. On Wednesday, China’s official Xinhua news service reported good news for Qualcomm and its local licensees:
Under the stimulation of China Telecom's large investment and procurement this year, the growth of China's CDMA cell phone market will be great, said an industrial insider. Competition will heat up as many GSM mobile phone manufacturers are entering the CDMA field, he added.

China's CMDA cell phone market rebounded in the end of 2008 despite gloomy market conditions. According to statistics released by market research firm Sino-MR, sales volume of CDMA handsets topped 1.29 million during December 2008, up 33.6 percent year on year and 183 percent month on month, marking a five-year high.
However, the report ended with a complaint about the royalty rates
Insiders with China Telecom consider China an important market for CDMA business and expect Qualcomm to realize the market condition and give more supportive pricing. Qualcomm refused to comment.
When in 2000 Irwin Jacobs negotiated government permission for CDMA to be used in China, Qualcomm granted Chinese manufacturers the best rates in the world — around 2.65% as opposed to 5% for other makers. (Qualcomm neither discloses nor confirms its royalty rates).

In fact, this deal has brought continued griping by Korean manufacturers paying list price. The Korea Times has run one-sided stories since 2004 airing selective leaking by the unhappy manufacturers. For example, it ran a March 2004 story (no longer on their website) entitled “Qualcomm's Royalty Policy Angers Korean Chip Makers”:
U.S. chipmaker Qualcomm’s royalty policy has angered Korean manufacturers of code division multiple access (CDMA) phones as it charges lower rates to China than to Korea.

A local CDMA phone maker contends that Qualcomm gives a “subsidy” to Chinese exporters as it collects as low as 5 percent of handset prices in royalties from them. But the rate is 5.75 percent for Korean makers.

Korean cell phone producers, which pay a fixed 5.75 percent for export and 5.25 percent for local sales to the San Diego-headquartered firm, shoulder higher royalty burdens in both domestic and global markets.

According to a contract between Qualcomm and one of China’s leading handset manufacturers, which The Korea Times obtained exclusively, the royalty rates are 2.65 percent for local sales and 7 percent for offshore shipments.

However, the contract says the export royalty will be cut to as low as 5 percent after the three years immediately following the license effective date.

The deduction ranges from 5 percent to 6.5 percent, depending on quarterly export volume, but the Chinese firm is entitled to enjoy the low rates “if more than 100,000 such subscriber units (handsets) are sold in the applicable calendar (year)” in overseas markets.
The KT report said that Chinese manufacturers signed a 10-year agreement in “early 2001,” which would mean these deals will be expiring in 2011.

A review of Qualcomm’s 2001 press releases reveals this May 21, 2001 announcement:
Qualcomm Incorporated … today announced that it has signed a commercial license with ZTE Corporation, a leading manufacturer of telecommunications equipment in China. Under the terms of the royalty-bearing license agreement, Qualcomm has granted ZTE a license under Qualcomm's CDMA patent portfolio to develop, manufacture and sell cdmaOne™ and third-generation (3G) CDMA2000 1x/1xEV network equipment. The license grants ZTE the right to use Qualcomm's technology and integrated circuits to make and sell cdmaOne and CDMA2000 equipment in China and worldwide. ZTE becomes the first company in the People's Republic of China to enter into a commercial license with Qualcomm.
Additional manufacturers signed licenses in July and November, with the latter including Huawei. Another 17 Chinese firms signed in January 2002.

The Xinhua story seems designed to help Chinese manufacturers (or China Telecom) pressure Qualcomm to lower royalties on the current deals, or at least when the licenses come up for renewal in two years.

Interestingly, there is no complaining about the GSM royalty rate, which is also secret. One report estimated the rate at 2-10%. However, data from the various Sendo-related lawsuits put the number more like 10-13%, suggesting that single-digit GSM royalties are only available to firms with big enough patent portfolios to negotiate a cross-license. Thus far, only a few Chinese manufacturers have enough patents to possibly put them in the latter category.

Thursday, February 5, 2009

MediaFlo hurt by DTV delay

The House voted yesterday to delay the Digital TV transition by nearly 4 months. A majority of both houses have decided that America is not ready for a transition and thus a delay to June 12 is warranted.

However, delaying the end of analog transmissions (and thus the surrender of those broadcast bands) is also delaying the deployment of new services by the communications companies who paid for rights to those bands, especially in the various 700 MHz bands.

Although AT&T and Verizon own such spectrum, the company most ready to use its spectrum is Qualcomm with its Media Flo service licensed by AT&T and Verizon.

Qualcomm issued a statement Wednesday (although it’s not on its website), with the most complete summary in Telephony magazine:
We are disappointed with the passage of legislation extending the DTV transition date to June 12th. … Due to the investments we made, we were ready for a February 17th transition to provide our innovative FLO TV service nationwide immediately. We are encouraged that several Congressmen and Senators who supported the delay stated that this would be a one-time delay only. In light of the fact that the legislation, as amended and finally passed by Congress, allows TV stations to transition voluntarily between now and June 12th, we cannot determine the specific impact of the final bill's passage on our MediaFLO business.
Due to interference its service causes around (analog) UHF channel 55, Qualcomm will have to delay rollouts in Boston, Houston, Miami and San Francisco; its efforts to have these markets exempted fell on deaf ears.

It’s hard to predict the commercial impact of the delay. Qualcomm has been aggressively planning this rollout for five years, and its speed to market was intended to preempt rivals (at least in the US) from gaining a foothold in the (presumed) lucrative mobile multicast TV market.

Sunday, February 1, 2009

Coding in space

Local telecom industry veterans, particularly those who got a good dose of information theory in their EE degrees, understand why Andy Viterbi won his Shannon Award. All would know that Viterbi developed what (proved to be) the maximum likelihood decoding scheme for convolutional codes, which remained the state of the art until turbo coding came along.

Most would know that every GSM cellphone (let alone CDMA cellphones) includes a Viterbi decoder. Some might know that one of Qualcomm’s first products (and I believe its first semiconductor product) was a Viterbi decoder chip.

A few would recall that Klein Gilhousen led the group that designed the LMP (Linkabit Microprocessor) which was used to implement the Viterbi algorithm for the Dual Modem.

What might not be known — I didn’t know until I researched the back story for the book — is that before cellphones, chips and the dual modem, the first real application for the Viterbi decoder was for NASA deep space probes.

In the late 1960s, NASA was planning two competing families of space probes for mankind’s first visit to the outer planets. The communication challenges were among the most daunting in history, where (due to the inverse square law) the signal strength a billion miles away was faint, and power (from nuclear decay batteries rather than solar power) is minimal.

Theoretically, the application of Claude Shannon’s ideas (and coding theory) could provide as much as a 10 dB signal gain. MIT-trained academics worked with two different divisions of NASA, both using coding theory:
  • Pioneer 10 and 11: Codex of Newton, Mass. worked with NASA Ames to apply Fano decoding
  • Voyager 1 and 2: Andrew Viterbi, Joe Odenwalder and others — first as a UCLA contract and then as a Linkabit contract — worked with JPL to use concatenated decoding and the Viterbi algorithm.
My research on this led to chapters on MIT and NASA for the book, as well as changes to other chapters. However, I’ve stopped quoting an ETA for the book given (as a part-time job) it’s coming very slowly.
In the meantime, two years ago I wrote a paper for the Journal of Management Studies, in a special issue on technology commercialization. The paper, entitled “Commercializing Open Science: Deep Space Communications as the Lead Market for Shannon Theory, 1960–73,” was published in December after several rounds of review.

Errata (Feb. 4, 2009). I received an email from Andrew Viterbi with a correction to the paper that will be applied to the book:
the first Linkabit customer was the Naval Electronics Laboratory Center (NELC) in San Diego… NASA was the second, but never became a large customer.
Most universities subscribe to the journal, so the paper is available as online from a university IP address (DOI: 10.1111/j.1467-6486.2008.00807.x). Normally I post a PDF of the pre-publication Word document, but there were major changes after submission during editing the page proofs. However, I would be glad to email the published PDF to anyone who asks (at ).

Graphic credit: Brewster Rocket comic from Sunday, February 1, 2009. Note to Tim Rickard: Pioneer 10 was launched first, but Voyager 1 was actually the first manmade object to leave the solar system and the most distant space probe. So your aliens would first see the Golden Record rather than the Carl Sagan memorial plaque.