Pez -- Speaking of SPWX, if you haven't seen it yet here are some recent comments from the software analyst at Salomon. Sorry for the length of the article, but there are a few interesting tidbits of info in here relating to speech recognition among all of the wireless commentary:
VC-CARRIER EVENT PROVIDES SNAPSHOT OF WHAT'S HOT/WHAT'S NOT IN NEXT-GEN
WIRELESS
Wednesday we attended a VC-carrier round-table event in the Valley focused on
financing, data-center penetration strategies, and "what's hot, what's not" in
next-gen wireless. Attendees included British Telecom, Deutsche Telekom,
Telecom Italia Mobile, and SK Telecom -- and a collection of venture players
trolling the waters for opportunities in chipsets, smart antennas, messaging,
Operations Support Systems (OSS), telco billing, and various flavors of
middleware. The dynamic was reminiscent of the recent CTIA show, with junior
vendors trying to penetrate the carrier data center for the all-important first
big deployment, while carriers ponder how to 1) offset eroding voice revenue
density with value-added data services; 2) be first to market with an
(obviously impossible) risk-free, pre-proven applications bundle; 3) improve
network management, provisioning granularity, and QoS, and 4) cut operating
costs. The following summarizes:
802.11 WLAN -- 2.5G PACKET WIRELESS INTEROPERABILITY
Given traction of high-speed wireless local-area networks (WLANs) in the
enterprise and the mass market via independent operators such as Boingo and
Sputnik, mainstream carriers are interested in software that will allow
subscribers to hop between 802.11 and 2.5G packet-wireless. Touted as a "3G
killer," 802.11 provides short-range wireless data speeds from 2 -- 54 MB/sec
(11 MB/sec on 2.4GHz for the "b" variant; 54MB/sec on 5 GHz for "a"). This is
illustrated by VoiceStream, which plans to link its GPRS network (rolling out
in late '02-03) with the 802.11b assets it acquired from MobileStar in January.
Initial service delivery will be via PC aircards, with handhelds to follow.
802.11/GPRS/CDMA 1x interoperability implies formidable software challenges in
network handoffs, voice integration, QoS, billing, and security that should
keep developers busy for years.
TCP/IP OVER WIRELESS REMAINS THORNY
Despite the much-bemoaned failure of early WAP services, the convergence of
wireless and the Internet remains a tectonic telecom trend. This is driven by
1) major content, portal, and media players eager to exploit the mobile
channel; 2) equipment and software providers such as Cisco, Sun, Oracle, and
Microsoft, and 3) enterprises seeking to leverage webified supply chain and
productivity tools. While it is conventional to dismiss the underlying
infrastructure as "collectively commoditized WAP gateways," there is much more
to the challenge of integrating voice and data, allocating bandwidth,
differentiated QoS, managing devices, IP addressing, security, and billing.
From the carriers' perspective, it remains early days in terms of delivering
TCP/IP over wireless.
CONFLICTING SIGNALS ON INFRASTRUCTURE VS. CONSUMER VS. ENTERPRISE
Venture-stage investors continue to wrestle with wireless software/services
strategies -- namely, is it best to target network infrastructure at the
carrier level, the consumer, or the enterprise? We confess to having come
full-circle at least twice in the past two years. The "arms dealer" appeal of
infrastructure is obvious. Messaging gateways or SIP-stack enabled
softswitches, for example, are going to be installed in the network regardless
of whether the branded portals, handset vendors, or carriers win the battle to
"own" the mobile subscriber. With the demise of advertizing-driven Web
business models, it has been fashionable to shun the mass market in favor of
carrier-centric strategies, but in the words of one operator, "at the end of
every device is a consumer." One of the most salient wireless data
developments thus far in 2002 has been the resurrection of "enterprise
enablement" -- a market tagged early to drive adoption, but which was slow to
bear fruit given the confluence of brittle circuit-switched networks and
cumbersome middleware. Not surprisingly, the beneficiaries have been
enterprise software majors like Microsoft, Oracle, and Siebel rather than the
alphabet soup of junior players that launched the first wave. Suffice to say
that the infrastructure-consumer-enterprise debate remains more lively than
ever among VCs and network operators.
UNSUNG HEROES: NETWORK MANAGEMENT SOFTWARE & OSS/BSS
We are strong believers that wireless data services introduce a multitude of
third-party content, hosting, messaging, Service Level, policy, and wireline
players into what has traditionally been a carrier-subscriber-voice minutes
relationship. That is, in order for the much-prophesied wireless Instant
Messaging, Unified Messaging, commerce, or premium alerts services to actually
function, there are suddenly multiple layers of switches, servers, routers,
multiplexers, and gateways involved that are totally beyond the control of the
wireless network operator. In this way, what was once a simple two-party
telephony relationship (or three with roaming) now begins to look much like an
e-commerce value chain, with most of the complexity actually on the wireline
side of the cloud. Similarly, IP-billing remains a major stumbling block for
the deployment of packet-data services, and the European operators at the event
(generally more advanced than the U.S.) pleaded to keep business models as
simple as possible -- noting that tiered-usage or value-based software models
might play well to investors, but simply cannot yet be reliably billed for. We
note that GSM-GPRS handoffs at the edge of metro areas are creating billing
headaches, as is basic prepaid SMS/data. Discourse on this subject continues
to be humbling.
MULTIMEDIA MESSAGING (MMS) -- INTERIM PRODUCT, NOT 3G ROI MESSIAH. ALL EYES ON
IM
Interestingly, there was much less discussion of MMS as an ARPU driver than
commonly heard in Europe (for example at the Cannes 3GSM show) -- perhaps
because simple SMS is just gaining traction in the U.S. We continue to regard
MMS hype as emblematic of the industry's desperation -- and believe that it is
an interim product bridging SMS and full-featured wireless Internet. The
probability that wireless picture postcards or audio clips will salvage the 3G
ROI proposition is remote. VCs and carriers at the event were much more
focused on wireless email and Instant Messaging, which leverages the U.S.' Web-
savvy demographic.
MULTIMODAL ACCESS -- WEAVING SPEECH RECOGNITION INTO THE FABRIC. COST-CUTTING
IS KEY
Speech recognition, as they say, is at the tip of everyone's tongue. We
believe speech is at the core of the converged network, along with multi-modal
access -- or blending voice and graphics (for example request a map or stock
chart verbally; receive it graphically).
Consistent the rest of our CommSoft coverage, we see telcos focused on boosting
margins through operating cost cuts. Thus, it is not surprising that speech
deployments dwell on "nuts and bolts" automated directory assistance and
customer care functions, rather than futuristic voice portal or unified
messaging "nice to haves."
While operators are keen to boost ARPU, they are hesitant to spend on next-gen
revenue creation tools, and cost control has assumed center stage amid a
general industry slowdown -- not to mention the liquidity jitters precipitated
by Sprint, Qwest, WorldCom, Williams, and Global Crossing. While capital
budgets capture the headlines in terms of large-cap equipment vendors, the real
money is operating costs, and the objective to boost margins by attacking
operating pain points. We regard network management software and speech
recognition as prime beneficiaries of this dynamic.
LOCATION BASED SERVICES, M-COMMERCE, BLUETOOTH -- THE SILENCE IS DEAFENING
At Wednesday's event (and consistent with our conversations with investors),
there was a singular lack of interest in location-based services, mobile
commerce, and Bluetooth. Much has changed in the last year. Carrier
representatives bemoaned the ineffectiveness of regulatory impetus for LBS
(through E-911 mandates) in spurring viable consumer services. M-commerce is
regarded as a fantasy in a U.S. market where SMS is just beginning to gain
traction -- and technically impossible prior to the widespread deployment of
2.5G packet wireless. Bluetooth appears to be a "solution in search of a
problem," and however complementary to WLANs as it may be, appears to have been
lost in the 802.11 buzz.
COMMSOFT STRUGGLING AMID TELECOM SHAKEOUT -- BUT IT'S THE SOLUTION, NOT THE
PROBLEM
CommSoft stocks have been punished amid the telecom shakeout, with our market-
capitalization weighted composite falling 29.2% year to date -- and no near-
term catalysts from wireless data, cable MSOs, wireline network management, or
enterprise IT spending. Our stocks are trading at 2-3x 2002E revenues and 1.3-
1.8x 2003E (excluding several negative-enterprise-value names). Nevertheless,
intelligence in the network continues to go "soft," and we believe that core
cost-cutting and revenue-creation tools in network management, speech
recognition, and wireless data are part of the solution, not the problem.
Taking a full-year 2002 view, we continue to favor Openwave Systems (OPWV, 1H,
$6) and Micromuse (MUSE, 1H, $9) -- followed by speech recognition names
SpeechWorks International (SPWX, 2S, $8) and Nuance Communications (NUAN, 2S,
$6).
News 
Market Data 
Discover 
