Semiconductor-laser market sustains steady growth

Semiconductor-laser market sustains steady growth

The total worldwide market for lasers is predicted to grow at approximately 17% per year for the foreseeable future

Jamie dreyfuss

mitsubishi electronics america Inc.

Driven by such new applications as video-on-demand, interactive learning, virtual reality shopping and full-motion video conferencing, the telecommunications market has been inundated with information about how this revolutionary commercial enterprise represents a tremendous business opportunity for aggressive opto-electronic companies.

Not so for fiber-optic laser components.

Mitsubishi`s analysis leads to the conclusion that the total worldwide market for lasers, including information superhighway construction, will continue to grow at approximately 17% per year for the foreseeable future. The picture is similar for the North American market segment. Although this growth rate may appear robust compared to the overall economy, a more relevant measure is against the rest of the semiconductor industry, which is growing at an estimated 20% per year. Components suppliers are bottom feeders on the great communications food chain.

Regulatory issues

According to market research concern Ryan, Hankin, Kent Inc. in So. San Francisco, the components market for public communications applications is highly dependent upon the regulatory environment in North America, specifically the actions of the United States government`s Federal Communications Commission. A small change in the regulatory environment can have a huge effect on technology vendors.

The current regulatory environment, for example, has created a high level of interest in the hybrid fiber/coaxial-cable architecture to provide broadband services from the regional Bell operating companies and multisystem operators. The popularity of the hybrid fiber/coaxial-cable architecture en courages laser vendors to continually develop new analog lasers. However, if through changes in the regulatory environment, regional Bell operating companies and multisystem operators decide to immediately pursue all digital networks, then system and component vendors will have to develop a completely different set of products. This is because digital networks use different components and architectures--for instance, hybrid fiber/coaxial-cable networks--than analog networks.

The case of high-definition TV exemplifies how seemingly small regulatory decisions have a significant impact on component vendors. By requiring television vendors and broadcasters to maintain compatibility with current video standards of the National Television Standards Committee, the regulations have effectively eliminated an analog high-definition TV market, which as proposed, was not backward-compatible with these standards. Thus, the FCC`s decision rendered obsolete all research and development already invested in analog high-definition TV components; vendors should continuously monitor the actions of the FCC and adjust projections of technical and business trends accordingly.

Market segments

The North American semiconductor laser market can be subdivided into segments by application. Two major groups of applications include fiber-optic and beam. End users of fiber-optic systems are typically either operating companies--such as the RBOCs, long-haul carriers, cable-TV multisystem operators and cellular providers--or private end users of local area networks or private branch exchange systems.

Beam applications include laser printers, bar-code scanners, laser pointers, CD-ROM drive/CD players, erasable optical disk drives, pumping for yttrium aluminum garnet lasers and measurement equipment. The combined North American market for all beam applications is relatively small (estimated at approximately $20 million), partly because most laser printer engines and CD players are made offshore. More importantly, however, the low price- typically $0.80 to $20-of the laser components used in these applications (lasers used in CD players have been less than $1 for years) affects market size The largest beam application in North America, on a dollar basis, is bar-code scanners; this market is estimated at approximately $10 million.

Fiber-optic systems are used in five major types of applications--access, interoffice/long haul/submarine (combined because the lasers used have similar characteristics), data communications, test equipment and other (cellular/plastic clad silica, military, research and development).

Access market--The access market is the telecommunications infrastructure from the central office or headend to the user, typically a residence. This portion of the market is associated with the information superhighway application. It has four application subcategories-feeder; hybrid fiber/coaxial-cable/cable TV; plain old telephone service/narrowcast/upstream; and digital fiber to the curb.

Feeder. For current installations, feeder is defined as the digital synchronous optical network-compatible link between the central office and a remote terminal or business user. Feeder demand is primarily driven by rebuild with some new construction.

For the most part, this is a captive market because many system vendors can supply their own lasers. Typically, 1300-nanometer Fabry-Perot lasers in either connectorized or pigtailed packages are used in feeder. Because the typical feeder data rate X distance product is less than 20 gigabits kilometers per second, today`s lasers are technically adequate and reliable. Therefore, there are no significant technical issues for laser improvement. Quantities of lasers for this segment are expected to remain relatively flat--40,000 pieces. However, the total revenue market is expected to contract because of price erosion (from $4 million in 1994 to $3 million in 1997).

Hybrid fiber/coaxial-cable/cable TV. The hybrid fiber/coaxial-cable/ cable-TV market is one of the largest non-captive markets for lasers.This segment is the portion of the broadband service network that delivers broadcast and specialty video to the home through a hybrid network of optical fiber and coaxial cable. For the purposes of this discussion, quantities of components used for cable-TV trunking applications in this segment have been included.

The hybrid fiber/coaxial-cable/cable architecture is primarily used by cable-TV multisystem operators; more than 90% of the component usage in 1994 was shipped in systems to multisystem operators. However, this is shifting because RBOCs are starting to invest in infrastructures that will allow them to provide video-distribution services in competition with the cable-TV multisystem operators. By 1997, it is expected that more than 30% of hybrid fiber/coaxial-cable demand will come from the RBOCs.

Until recently, the RBOCs were forbidden to provide cable-TV services. However, in 1993, the FCC issued a ruling that allowed the RBOCs to provide video dial tone. The RBOCs, led by Pacific Bell, are beginning to install hybrid fiber/coaxial-cable systems to compete with the cable-TV multisystem operators in their serving areas.

The 1300-nm distributed-feedback laser is the workhorse of this market because it can meet the transmission requirements (80+ video channels per laser) at the lowest price versus other technologies. The average selling price of lasers for cable TV is $2500 to $5000 each, depending on the specification and quantity.

The price for analog distributed-feedback lasers is under pressure because of competition in this market, while the analog specifications (linearity, usable power) continue to improve. The trend for laser manufacturers is to have the capability of providing 110-channel testing while still meeting customers` linearity requirements.

The major system manufacturers of hybrid fiber/coaxial-cable/cable-TV systems are AT&T Network Cable Systems, Scientific-Atlanta Inc., General Instruments, Philips Broadband Networks Inc. and ADC Telecommunications Inc./American Lightwave Systems Inc.

Higher bandwidth requirements will drive the need to reduce node size, and this will result in an increase in the required number of lasers. However, the recent high growth rates for components will not continue because the market for downstream cable-TV lasers is reaching saturation.

Plain old telephone service/nar rowcast/upstream. This segment is defined as the portion of the hybrid fiber/coaxial-cable architecture that provides two-way communications, including telephony. While the laser component technology (distributed feedback versus Fabry-Perot) has not yet been decided, customers are currently leaning toward the use of distributed feedback. This choice will be driven by a number of factors, including upstream bandwidth requirements, laser reliability and cost/performance tradeoffs.

In the current regulatory environment, many service providers and system vendors believe this portion of the hybrid fiber/coaxial-cable architecture is the most cost-effective method to provide full-service, two-way communications (including new interactive services) to the home. However, recent developments based on RBOC field trials indicate that alternative architectures (digital) appear promising. Nevertheless, it appears this segment will provide laser vendors with good market growth in units and dollars.

The RBOCs will be installing complete hybrid fiber/coaxial-cable systems because they do not have coaxial networks in place. The multisystem operators will be installing both complete systems (new builds) and the upstream portion. Therefore, growth in multisystem operator component use is much larger for the upstream portion of the system than for the downstream.

Digital fiber to the curb: The narrowband fiber-to-the-curb architecture has been replaced as the system design of choice for delivering telephony to the home by hybrid fiber/coaxial cable. However, broadband digital systems are expected to be developed to compete with, and eventually replace, analog-based hybrid fiber/coaxial-cable systems.

In the long term, the most efficient way to deliver broadband services is to have one infrastructure using switched digital technology (asynchronous transfer mode-based). This expectation is based on past trends in other communications areas where application of digital technology eventually becomes the lowest cost solution. Meeting the reliability and system power requirements (for example, the provision that telephone service continue through a power blackout) adds a significant share to the system cost for a full-service network. Until these issues are resolved, experts will differ on which technology (hybrid fiber/coaxial cable or digital fiber to the curb) will be the most cost effective.

Based on these concerns, the RBOCs will continue to advance all-digital fiber-to-the-curb development through limited deployment. It is estimated that in 1997, the RBOCs will install digital fiber-to-the-curb network architecture to less than 1 million homes in North America, using approximately 60,000 1300-nm FP lasers. Manufacturers of digital broadband fiber-to-the-curb systems include Broadband Technologies, AT&T and Raynet Corp.

Interoffice/long haul--The inter office/ long-haul system is a Sonet-compatible system that serves central office-to-central office and long-distance trunking applications with data rates from OC-12 to OC-192. The same system equipment, component technology and data rates are required for central office-to-central office and long-distance communications; thus, they are considered a single component market. By 1997, the broadband digital services to businesses will also use this same equipment; therefore, a small part of the information superhighway demand is included in this segment.

Traditionally, the long-haul market is the major market for lasers. This most captive market (1994 usage is approximately 40% non-captive) is primarily for OC-48 systems. These OC-48 systems use 1300- or 1550-nm distributed-feedback lasers that operate at 2.5 gigabits per second. Although OC-48 systems now dominate this segment, most of the laser component revenue growth in this segment through 1997 will come from the introduction of OC-192 systems.

Two types of systems for OC-192 data rates are being proposed--a single externally modulated 1550-nm distributed-feedback laser at 10 gigabits per second with erbium-doped fiber-amplifier repeaters; and four wavelengths of 1550-nm lasers externally modulated at 2.5 gigabits per second, each with erbium-doped fiber-amplifier repeaters.

First field trial installations will occur in 1995. System vendors are willing to consider any source for OC-192 components because they are not readily available. However, captive suppliers are expected to enjoy a dominant market share by 1997. The market is expected to move away from multiple-wavelength solutions by that time because single-wavelength solutions will cost less.

Demand for higher speed and performance long-haul systems has grown steadily for more than a decade, and this trend is expected to continue. The majority of these systems will continue to use internally supplied components. However, this market consistently remains attractive for component vendors that have leading edge technology. Systems vendors are AT&T, Northern Telecom Ltd., Alcatel Network Systems Inc., Fujitsu Network Transmission Systems Inc. and NEC Corp.

Submarine--AT&T is the only North American customer in this market. Because this is mostly a captive market, it is difficult to estimate. Typically, 1480-nm pump lasers are used for undersea repeaters, amplifying a 1550-nm source.

Data communications--Fiber-optic data-communications applications tend to use either transceiver products or optical-link cards at 10 to 266 megabits per second, typically using light-emitting diode-based products. These applications are price-sensitive because distances tend to be short enough for copper solutions to compete. As a result, the laser-based market for data communications (private branch exchange and LANs) is relatively small.

The expected areas of largest growth are in Fibre Channel Standard (using 780-nm FP lasers from Mitsubishi) and ATM (using 1300-nm FP lasers).

Test--The primary market for lasers in the category of test equipment is for optical time-domain reflectometers. OTDRs use high-pulsed power FP lasers. As more fiber is installed, more low-cost field usable test equipment will be required. However, while demand for lasers in this market will increase, revenues are expected to remain flat.

Other (analog cellular/personal communication services, etc.)--For now, a relatively small, but promising, portion of the laser market is for systems that connect cellular/PCS base stations. Analog lasers (1300-nm distributed-feedback lasers) are required to carry 900-megahert¥or higher carrier frequencies.

There is no windfall in growth expected for the laser component community as a result of the large communications infrastructure project known as the information superhighway. But component suppliers that have strong product technology and good market share can expect to continue to grow at a reasonable rate in the communications market of tomorrow. u

Jamie Dreyfuss is the high-frequency device product manager at Mitsubishi Electronics America Inc., Sunnyvale, CA.

More in Market Research