KDDI 3Q Rock'in Results and BREW Comes to a Boil
KDDI 3Q Rock'in Results and BREW Comes to a Boil

KDDI 3Q Rock'in Results and BREW Comes to a Boil

KDDI 3Q Rock'in Results and BREW Comes to a Boil

Berman warns, however, that with a typical handset cost of 37,000 yen in the latest quarter, “it will require about 23 months to cover [the] cost for upgrade customers.” He goes on to state that this creates a difficult situation for KDDI, wherein there’s lots of room for concern for the cost of the 2G-to-3G upgrade cycle. Remember, too, that KDDI has committed to launching a 1X EV-DO upgrade in late 2003. Will it have to commit to another two-year cycle to convert subscribers to the new system? If so, the expense half of their balance sheet may get a little out of control no matter how strong the revenue side remains.

It’s another one of those weeks: there’s so much happening on Japan’s mobile Internet, I’m not too sure what to cover first. Let’s start with KDDI’s 3Q results, released to analysts last week.

CSFB sr. telecoms analyst Mark Berman put out a note to investors in which he stated that KDDI’s CDMA 1X (3G) subscriber ARPU fell 4 percent in Q3 (September to December 2002) versus Q2 – ending up at 9,240 yen. As a result, 3G ARPU was 21 percent higher than KDDI’s Au average (3G plus 2G), down from 25 percent in Q2 and 38 percent in Q1. “The 1X user is effectively spending about 1,600 yen per month more than KDDI’s au division average,” said Berman.

It looks like this positive difference is trending down – a phenomenon that I’ve seen here before; initial high usage rates of new technology tend to drop off over time (this happened with Java). Nonetheless, the 3G ARPU results are pretty darn good and the carrier must be pleased. This makes DoCoMo’s FOMA results look decidedly weak (see “KDDI Earns $256.5 Million on Popularity of 3G Service).

Berman warns, however, that with a typical handset cost of 37,000 yen in the latest quarter (down from 42,000 yen in the first half), “it will require about 23 months to cover this cost for upgrade customers.” He goes on to state that this creates a difficult situation for KDDI, wherein there’s lots of room for concern for the cost of the 2G-to-3G upgrade cycle. Remember, too, that KDDI has committed to launching a 1X EV-DO upgrade in late 2003. Will it have to commit to another two-year cycle to convert subscribers to the new system? If so, the expense half of their balance sheet may get a little out of control no matter how strong the revenue side remains.

To be fair, President Onodera has said that the 1X EV-DO upgrade will be limited, but, as Berman points out, they will still have to expend capex funds to create the service (and I hear it’s a much more expensive upgrade than the switch from cdmaOne to CDMA 1X is). Berman ended by maintaining his NEUTRAL rating for investors.

In other KDDI news, the carrier issued a statement at the end of January outlining its BREW (Binary Runtime Environment for Wireless) plans; the carrier had previously committed to a roll-out before the end of March 2003 (that would be next month). BREW is Qualcomm’s answer to Sun’s Java, and provides a similar onboard execution environment, i.e., the cell phone can download and run miniprograms, although there are significant differences in how the two are implemented.

Basically, Java is open and it costs third-party developers nothing to create mobile Java applications, but the carrier assumes the not-insignificant burden of creating a provisioning system (to allow handsets to access and receive Java apps downloaded over the air interface), and also has to worry about testing, security, billing, etc. And there’s no guarantee that the developer will earn any revenue. BREW solves most of these problems for both the carrier and the developer, but both have to pay Qualcomm ($$), and the company inserts itself into the carrier-subscriber relationship as content source – something that I can’t imagine any Japanese carrier going along with. I guess this will be modified for Japan; in a later Viewpoint, I will present extra information on how the system will work (especially how the billing will operate and how developers will have their BREW apps tested and certified). In the meantime, if everything in the press release comes to pass, it will be a pretty impressive system.

The first BREW phone capable of over-the-air download, Toshiba’s A5304T, is scheduled for release later this month, and three applications will be preinstalled: NAVITIME (navigation app that provides map information showing info such as traffic jams and walking routes), Heart Mail (for sending mail containing pictograms and animations), and Team Factory (enabling group members to immediately confirm the current status of friends or associates). I think all of these already exist as Java versions or as preinstalled BREW apps (available on KDDI’s first, none-downloading, BREW phone released last spring). Significantly, IBM Japan has teamed up with KDDI to offer mobile WebSphere Everyplace access, among other services.

All in all, this looks pretty interesting, but I’ll be interested to find out more details. Also, Java has a serious head start and a significant chunk of developer mindshare; 2003 is shaping up to be even more interesting than ’02.

Finally, I was asked earlier today for my opinion on why camera phones took off as they did in Japan.

My response was along the lines that a camera-on-a-phone makes for a perfect combination of a fun, social activity (taking snaps when out with friends at a bar or pub) with a way to send them off to absent pals (those still stuck at the office due to an Evil Boss, say). I worked at Fuji Xerox for a couple of years in the middle 1990s, and we often went out for office get-togethers or Happy Hour on Friday evenings. There was always a raft of (print) photos to look at on the following Monday morning (“Oh look, Taka-kun was trying to steal kisses from Furu-chan – and look how drunk Yuki got!!”).

Now consider this real-life, human activity in a newer era when everyone has a cell phone – it’s not too difficult to see why camera keitais took off as fun consumer toys. And, as we reported in WWJ video episode #46, NTT DoCoMo is due to launch the first million-pixel-class CCD camera phone in the first half this year, and that should, I think, help boost commercial and non-play applications.

— Daniel Scuka