TMF Post: Reconstructing Qualcomm’s Earnings
boards.fool.com
A lot of confusion seems to be surrounding this quarter’s results, especially judging from the amazingly uninformed commentary from the popular media talking heads (e.g. CNBC the morning following earnings release). True, there was a change in royalty accrual, which throws off the numbers, especially reported GAAP numbers, but the bottom line is quite simple. This company is still firing on all cylinders, and has had an exceedingly strong quarter and fiscal year, in absolute terms as well as relative to the expectations they had set for themselves the same last year.
Fortunately, the company management has gone out of their way to provide the numbers in a variety of flavors to allow comparability from one time period to the next, providing QTL (the technology licensing division that tallies up the royalties) revenues using the “previous method” and the “new method” of royalty recognition, for both pro-forma (excluding QSI, the strategic investment division, thought to be not directly relevant to the company’s core business operations), as well as including QSI—all of which are yet different from actual GAAP numbers. To place it all in perspective, the company has made available a schedule that provides a Comparison of the "Prior Method of Estimating Royalties" and the "New Method of of Estimating Royalties" (why do they make it sound like it’s a patent application?), at:
qualcomm.com
When it comes right down to it, the difference between what the prior method (which books estimated royalties when accrued, and is therefore typically at some variance from the amounts ultimately received), and the new method (which books royalty payments when actually received) would have produced for QTL revenues for the current quarter is minuscule—less then $2M in revenue out of about $402M for the quarter. So we can just as well ignore it. It will make a slight dent in comparability of the December numbers, which would have otherwise included the estimated spike associated with Christmas sales—these would now have to wait until the March quarter of FY 05 to show up in the numbers.
The accounting change has no bearing whatsoever on the business, and my discussion will focus on the latter. First, for the QCT (chipset) and QIS (information services) segments, the numbers are totally comparable, and for QTL I will use the “previous method” numbers—so I can compare them to the time series I have been tracking for the past few years.
Before looking at these numbers, my reading of the earnings release (available as usual at qualcomm.com ), and subsequent listen to the conference call replay (http://www.qualcomm.com/ir/webcast/earnings_q4_04.html) was seeking answers to the following key questions:
1. What is the impact of the accounting change and how will it affect our ability to assess performance?
2. The WCDMA factor: handset (and royalty) growth prospects, and relative position of QCT customers.
3. QCT competition, selling prices, customer wins for WCDMA
4. Is plain vanilla CDMA slowing down?
5. The new strategic ventures: what’s with Media FLO, and the Iridigm acquisition?
I have already answered item 1, at least for the purpose of this quarter. I will try to share what I have gleaned and what I am speculating about the other items through the rest of the discussion.
OVERALL NUMBERS
Before delving into the segments, it is important to see the big picture, and just how impressive the company’s performance has been this past quarter, and indeed this entire past year.
Qualcomm has crossed an important threshold. It is now a $ 5 Billion company in annual revenues. Using pro-forma numbers (excluding QSI) with the “previous method” of royalty accrual, for comparability, the table below shows that annual revenues grew from $3.8B in FY03 to 5.1B in FY04, for a 33.4% annual growth rate. Earnings (pre-tax) correspondingly grew 48.1%, with nearly half of revenues going to the pre-tax bottom line (we can see that the pre-tax margin expanded from 45% in FY03 to 50% in FY04, while the after-tax margin similarly increased from 30% to 36%).
REVENUES EARNINGS pre-tax EARNINGS after-tax CASH on hand
FY 03 3846 1733 1161 5400
FY 04 5131 2566 1833 7600
Growth 1285 833 672 2200
% growth 33.4 48.1 57.9 40.7
Qualcomm has the cash in the bank to show for its performance. Its cash and marketable securities position increased an astounding 40.7%, to $7.6B—this is about 1.5 times annual revenues! At the same time, the company tripled its dividend to shareholders, and Dr. J. in the CC promised more along this line.
Its year over year increase in revenues for the quarter (Q4 04 vs. Q4 03) is 57.3%, with a corresponding increase of 87.9% in earnings—while the margin increased from 40% to 48%.
What they told us last year. To get the full measure of how strongly this company has performed, compare these numbers against the guidance the company management provide at the same time last year (at the time of the Q4 03 earnings release):
Based on the current business outlook, we anticipate that revenues excluding the QSI segment will grow by approximately 5-9 percent year-over-year and earnings per share excluding the QSI segment to be in the range of $1.37-$1.43 for fiscal 2004, compared to $1.42 last fiscal year.
Instead of 5-9 percent, they delivered growth of 33.4%, and the EPS came in at $2.18 per share (adjusting for the intervening stock split) instead of $1.37-1.43.
They also said, then:
We estimate the CDMA phone market to be 131-136 million units in calendar 2004, and we estimate a decrease of approximately 8 percent in average selling prices of CDMA phones for fiscal 2004, upon which royalties are calculated, compared to an estimated decrease in fiscal 2003 of 2 percent.
Of course, the CY 2004 best estimate for CDMA handsets is at now 170M, while the ASP has increased, and is again projected to increase as more WCDMA and high end models form a greater part of the mix.
And for some perspective: having recently looked at Nokia’s numbers (http://boards.fool.com/Message.asp?mid=21457886 ), Qualcomm’s performance becomes even more impressive when considered against the backdrop of market trends reflected in the performance of the largest handset maker. The divergence in relative performance between these two companies is striking (granted, they are not strictly in the same business, though they live and die by the growth in overall market for cell phones). Nokia posted a 20% year over year decline in overall operating profit for the quarter, and 44% decline in operating profit in its mobile phone division (offset partly by increase in multimedia)—compare this to the kind of explosive growth posted by Qualcomm. It is rather ironic that Nokia’s results were greeted with such enthusiasm by analysts and by the market, while Qualcomm’s numbers have so far apparently caused concern.
SANJAY’s CHIPSET BUSINESS
Below are QCT quarterly revenues and earnings (in $1000’s), going back to Q1 2000.
QCT
Revenues Earnings margin (%)
Q100 352,395 127,690 36
Q200 279,186 89,977 32
Q300 338,132 109,573 32
Q400 268,989 64,281 24
Q101 330,632 84,180 25
Q201 364,059 84,866 23
Q301 333,115 70,582 21
Q401 336,881 65,917 20
Q102 359,144 86,941 24
Q202 343,815 77,724 23
Q302 404,253 117,524 29
Q402 483,617 158,334 33
Q103 709,681 288,282 41
Q203 652,873 223,520 34
Q303 557,240 163,114 29
Q403 504,500 121,808 24
Q104 751,818 260,661 35
Q204 711,257 257,956 36
Q304 789,978 254,160 32
Q404 845,000 271,000 32
The strength in the core chipset segment of the business is simply remarkable. QCT boss Sanjay Jha indicated that they shipped a record 39M MSM (subscriber unit) chipsets this past quarter—this is nearly twice as many as they shipped in the same quarter last year (20M). He also indicated that they continued to be supply-constrained, i.e. unable to meet customer demand because of foundry capacity. He indicated very good visibility for the coming quarter, which he expects to remain at about the same level—though it would go up to the mid 40’s were it not for the capacity issues. To address the latter, he indicated signing a capacity-reservation agreement with IBM, which he expects would alleviate the current crunch.
I ran the usual 3-quarter moving average for QCT revenues (i.e. the average of the current quarter with the last two preceding ones), to smooth out quarterly fluctuations. Here it is in Table form.
Q300 323,238
Q400 295,436
Q101 312,584
Q201 321,227
Q301 342,602
Q401 344,685
Q102 343,047
Q202 346,613
Q302 369,071
Q402 410,562
Q103 532,517
Q203 615,390
Q303 639,931
Q403 571,538
Q104 604,519
Q204 655,858
Q304 751,018
Q404 782,078
Sanjay also indicated particular strength in infrastructure chipsets, especially for EVDO in the US. When asked about WCDMA customer wins, he indicated they now had 26 handset manufacturers who had orders with QCT, including three of the top six makers (LG, Samsung and Siemens). There was no discussion of market share expectations—we are clearly nowhere near the desired 50% target...
Expect more investment in the coming year, to address the dizzying array of wireless broadband technologies, multimedia capabilities (gold and platinum EVDO multicast which combine elements of OFDM with EVDO), HSDPA, FLO (a new proprietary air interface technology, optimized to provide maximum efficiency on a Forward Link Only basis, for multicasting streamed multimedia and clip downloads to handsets)…This should keep margins at about their current 32% level.
ROYAL ROYALTIES
Using the “previous method” of royalty accrual, the time series of quarterly revenues and earnings for the QTL revenues are shown below:
QTL
Revenues Earnings margin (%)
Q100 183,845 168,890 92
Q200 157,197 139,968 89
Q300 161,301 139,440 86
Q400 165,568 147,466 89
Q101 186,824 174,139 93
Q201 225,646 206,663 92
Q301 180,129 152,890 85
Q401 189,340 172,102 91
Q102 210,803 188,688 90
Q202 193,955 171,535 88
Q302 198,853 174,450 88
Q402 243,481 221,500 91
Q103 255,423 229,409 90
Q203 260,110 236,192 91
Q303 242,479 218,363 90
Q403 242,184 212,657 88
Q104 353,421 324,673 92
Q204 390,257 361,591 93
Q304 436,449 398,187 91
Q404 402,000 362,000 90
With WCDMA royalties constituting a growing fraction of the total (26% estimated this quarter), the limitation of the “previous method” becomes more evident, as more of the total is contributed by manufacturers who do buy their chipsets from other suppliers—unlike CDMA 2000. In last quarter’s discussion of results, I stated that I would not be surprised to see the 05 number (of estimated WCDMA handsets) close to 50 or 60M. Sure enough, the company’s guidance in this CC estimates WCDMA shipments in CY 05 at 55M. With about 165M CDMA handsets forecast for CY 05, that puts the total at about 220M, an increase of about 30% from CY 04. Given the company’s usual caution, and the likely strength in the handset market driven by a plethora of new handset models and enhanced capabilities, we may well exceed that estimate yet again.
Without WCDMA, the estimated CDMA 2000 handset total increase in CY 05 (over CY 04) is about 20M units, or about 13% year over year. That is relatively modest, and includes EVDO handsets, which are expected to grow at a faster rate as US networks are deployed. Plain vanilla CDMA, if there is such a thing left, is projected to remain flat or increase only slightly; EVDO on the CDMA side, and WCDA are expected to provide the growth engine for the coming year.
Below is the 3-quarter moving average of the QTL earnings series— this may be the last quarter I present these numbers, as the company stops presenting numbers using the “previous method”. While they have already provided adjusted past year numbers using the “new method”, I’m not sure it’s worth the trouble to rebuild my spreadsheets.
Q300 167,448
Q400 161,355
Q101 171,231
Q201 192,679
Q301 197,533
Q401 198,372
Q102 193,424
Q202 198,033
Q302 201,204
Q402 212,096
Q103 232,586
Q203 253,005
Q303 252,671
Q403 248,258
Q104 279,361
Q204 328,621
Q304 393,376
Q404 409,569
In addition to the growing WCDMA contribution to the QTL numbers and overall bottom line, royalties at the end of FY05 and beyond will benefit from the end of contractual agreements whereby certain CDMA royalty streams have been shared with “two other entities” (believed to be Motorola and Lucent). Bill Keitel has estimated the impact of that on QTL to be around $320M-$360M in extra revenues in FY 05—at the magical 90% margin, that’s nearly $300M in profit, more than the entire QCT operation!
STRATEGIC DIRECTIONS IN WIRELESS AND INTERNET SERVICES
In my previous discussions, I had not given much attention to the QWI segment, which brings together such disparate services as the Omnitracs truck and fleet asset tracking services, and the BREW content download service. Part of the reason is that the contribution of this segment to the bottom line was just not that significant; in fact for a few years, it was a net money loser—largely because of the developmental cost of BREW. In checking their numbers this past quarter, I was pleased to see that segment is now contributing $163M to this quarter’s revenues, an increase of 11% from Q403, of which $16M went to the bottom line, a doubling of the $8M contributed to Q403. For the year, QWI contributed a respectable $586M to revenues, and $30M to earnings (increases of 17% and 11%, respectively, over FY03). It is good to see the investment in BREW finally paying off, with about 37 commercial operators now relying on it for data content on their 3G networks. We may well be at the point where growth of BREW usage and revenues may snowball, as the number of BREW-enabled handsets in the population has gone from 14M in Oct 03 to 28M in Oct 04, and the number of downloads has more than tripled (from 55M in 03 to over 180M in 04).
Building on the success of the BREW business model, Qualcomm’s Paul Jacobs appears emboldened to embark on a major new strategic initiative in the distribution of content to wireless handset users. The recently announced MediaFLO venture will indeed build out a network in the 700MHz spectrum, and use it to efficiently distribute multimedia content, especially TV shows, in either live streaming mode or in the form of downloadable clips (“clipcast”), using the new Forward Link Only air interface. The entity will then serve as a wholesaler of content to carriers, who will distribute it to their retail customers. Paul Jacobs is convinced that the future is in mass distribution of multimedia, especially video content to wireless devices. To find out if he is correct, Qualcomm and possibly other investors will invest about $800M over 4 or 5 years. There were several questions during the CC about the business model, and it appears at this stage to be of the “build it and they will come” variety. We’ll probably dedicate another thread to discuss MediaFLO—is it another wireless knowledge or Globalstar, or will it indeed revolutionize the way people receive, watch and interact with multimedia content?
This space is definitely worth watching. Just when you think the company has reached a certain plateau of sorts in terms of the business it has mastered, it launches a new bold concept and venture aimed at creating a growing income stream in concert with the further development and deployment of its wireless technologies.
In summary, the last quarter and past fiscal year have seen outstanding performance by this company, well in excess of repeatedly escalating expectations. All indications are that the company remains on top of its game, and is deepening its essential role not only with its traditional core of CDMA carriers, but also in the growing realm of WCDMA network operations. There is every reason to believe that the company’s strategy will handsomely pay off in the coming year in terms of its growing role and significance globally, as well as in terms of shareholder returns. I remain convinced that there is no better-placed company to benefit from the continuing growth in the deployment and adoption of 3G systems and services around the world.
Opinions and thoughts are always appreciated.
BRational |
