By Theresa Foley

Satellite manufacturing capacity around the world is about twice the amount required to meet demand for the next few years. Demand is pegged at about 25 to 30 geostationary satellites per year, presenting the five big manufacturers that build most of the world’s satellites with the difficult task of winning enough business to stay in the game.

Compounding the oversupply problem is fierce competition on price and single-digit profit margins that are far lower than managers and investors prefer. Wall Street tends to view manufacturing as somewhat of a dead-end, useful only for its ability to give a company entrée into more lucrative service ventures. Despite the predicted boom in satellite services, there will be no riding on operators’ coattails for satellite manufacturers this year.

Orders for new satellites in 1999 appear to be slightly below the norm. Approximately 25 geostationary communications satellites were ordered from the five top manufacturers, and not all of those are identifiable as firm, funded contracts with established customers. According to Matra Marconi Space’s director of telecommunications systems, Jean Michel Aubertin, only seven of those contracts were competitively procured with an open process, while the rest were “captive contracts” due to a relationship between customer and manufacturer. “There is a lot of pressure on cost,” he says. “Sometimes very aggressive prices are proposed to operators just to fill the factories.”

The biggest surprise may be the emergence of Lockheed Martin Commercial Space Systems (LMCSS), boosted by big orders from GE Americom and its own Astrolink unit, in the leading position for new orders in 1999. LMCSS won nine orders, including an order for four satellites from its sister Astrolink company; Space Systems/Loral (SS/L) got six, including at least one from a Loral-owned company; Hughes Space and Communications (HSC) received five orders, including at least two from its own units; Alcatel got one; and Matra, one.

Tracking only the prime contract awards presents an incomplete picture, however, since the companies also win orders for non-telecom satellites that are built in the same factory, and equipment or payload subcontracts generate more business as well.

Profitability is one yardstick by which the satellite builders are judged. Financial analysts say the Hughes unit acquired by Boeing in January, with its economies of scale from making so many satellites, has the highest EBITDA margins in the industry. The other manufacturers appear to have EBITDAs in the ballpark of nine percent or lower. An illustration of just how small a portion of company value is represented by manufacturing, in investors’ eyes, is a report last year from Banc of America Securities satellite analyst Armand Musey. It valued Loral’s satellite manufacturing business at only $1.92 of the $34 projected share value of the company’s stock at year-end 2000, and the Hughes/Boeing unit at $5.23 of the projected $75 stock value.

The Issue Of Reliability

The satellite builders have yet to shake completely the negative image problem that developed after a series of highly visible hardware failures starting in 1997. While the track record appeared to have improved in 1999, the perception remains that a serious, ongoing reliability problem plagues satellites.

Roger Rusch, president of consulting company Telastra of Palos Verdes, CA, has been tracking satellite failures over several decades, gathering data on more than 1,000 failure incidents. Rusch says the share of failures tracks closely with a company’s market share, with few exceptions. Companies are loath to disclose details on the failures and take steps to minimize release of information on the specifics, he adds. His conclusion: “No contractor is perfect, and some are better than others.”

The story from the manufacturers is that their products are reliable. In a reliability analysis from one of the companies, which asked not to be identified by name, the conclusion was that historically, more transponder service–40 percent more to be exact–has been provided than was originally contracted for by customers. The analysis shows Loral to have the most experience in providing transponders aboard three-axis satellites, followed by Hughes and Lockheed. Loral also has the highest transponder completion percentage for missions since the 1980s, according to this analysis, which shows a 93.8 percent rate of successful transponder operation for SS/L satellites, versus 92.5 percent for Hughes’ three-axis and 92.7 percent for Hughes’ spinners.

The prime contractors place part of the blame for hardware problems on their suppliers, although the big manufacturers will not discuss specifics about who is at fault. Several builders have imposed new processes to correct the deficiencies in how the suppliers were being managed. The most visible evidence of defective components getting through the quality control system was a series of flawed TWTAs (travelling wave tube amplifiers) that delayed several satellites being built by Space Systems/Loral and Lockheed Martin in 1999.

A separate series of parts failures started in August 1999 when Matra discovered solar array problems on the NSS KTV satellite it was building for New Skies Satellites and took the step of returning the satellite from the launch site in Kourou back to the factory in France for repairs. Similar solar cell defects were reported on other Matra and Alcatel satellites prior to launch.

“Industry had gotten complacent about managing subcontractors,” says Alexis Livanos, executive vice president of operations for SS/L. The result was that trivial changes were being made to components without properly being reported. John Klineberg, president of SS/L, says subcontractor quality control has been tightened following the TWTA problems.

Hughes had well-publicized problems with the spacecraft control processors on several orbiting HS 601s, resulting in one satellite failure. The trouble was traced to a tin- plating process on electronic parts. Hughes also has new management and supplier control processes into place. “Some problems in the schedule for satellite production in 1999 were caused by supplier quality. We now have a great deal of emphasis on enhanced support,” Randy Brinkley, HSC senior vice president of programs, says. Hughes also is trying to build strategic relationships with certain suppliers, although Brinkley would not be more specific on the supplier names or terms of the arrangements.

Questions have been raised about whether the reduction in production time, which has leveled off somewhere between one and two years, has led to some of the failures, but so far no direct connection has been made. A buyer will not find much uniformity among the various builders and models regarding the number of months required to build a satellite. Hughes can build an HS 376 in 13 months, while an HS 601 takes 20 months and an HS 702 takes 28. Alcatel requires 24 months for a new satellite, although that can be cut if a customer buys an already-built or partially-completed satellite. Matra Marconi Space aims for a 27 to 32 month production cycle, but has been able to deliver satellites in as little as 18 months.

The builders say a new emphasis is being placed on testing satellites during these compressed schedules as a result of the failures that have hurt the industry in the last two years of the 1990s. And they continue to claim high reliability, based on their own statistics, to counter the customer and insurer perspectives that satellite reliability is far lower than desired.

“Building the satellites faster may be better for the customers, but the manufacturers may have created a new problem for themselves by increasing industry supply without creating a corresponding increase in demand,” Musey says, which is part of the reason behind the partially filled factories.

The Next Generation Of Satellites

In any case, the future does not appear to hold much change in terms of sheer numbers of new satellites. A steady and decidedly flat pace of 22 to 30 commercial orders per year or less is assumed through the next decade. But even if the number of satellites built is static, revenues for manufacturing may go up anyway, since satellites are getting more expensive.

Total satellite manufacturing revenues were $11 billion in 1998, and will climb to only $11.3 billion in 2000, a “sleepy” growth rate of only four percent a year, Musey’s report says. The technology on board improves every few years, allowing manufacturers to charge more and operators to serve a greatly expanded number of users with the same number of satellites.

“The power and class of satellites has been doubling every four to five years,” says Dan Collins, SS/L vice president of worldwide marketing and sales.

“Operators are looking to squeeze every dime out of the cost base. The quickest way to lower the cost per transponder on orbit is with bigger satellites,” adds Dan Ozley, vice president of marketing and sales at LMCSS.

More powerful satellite models are being slowly introduced. Hughes is bringing out upgraded versions of the HS 702 and 601. The 702 will have up to 25 kW power and 100 transponders, nearly twice the 601’s 10 kW and 60 transponders. SS/L is working on the 20.20 satellite design, announced a year ago, but progressing slowly. Klineberg says SS/L has no orders for the 20.20 yet and has slowed the project down. SS/L also is offering a stretched version of the FS1300 bus to go head-to-head in competition with the Hughes 702.

Lockheed Martin’s strategy is to grow its A2100 platform to a 5,000 kg, 12 to 22 kW payload power version, which will compete in the bigger class, rather than coming up with a new design. Alcatel is developing the Spacebus 4000 series, with a 10 to 20 kW power range, which also has yet to win a first customer. Pierre de Bayser, Alcatel Space senior vice president for sales and marketing, said the first 4000 would be delivered in 24 months, assuming one of several customers to whom Alcatel has bid the model decides to buy it. Matra is designing the Eurostar 3000 as its most advanced platform, with a 10 to 15 year lifetime, capable of carrying payloads weighing up to 1,000 kg that use 6 to 12 kW of power. The 60-transponder satellite should be available around 2001, but as of early 2000, Matra was still trying to find its first buyer for the new satellite. The company hopes to sign its first contract for a 3000 this year.

Intelsat, New Skies, Echostar and GE Americom are among the operators currently shopping for new satellites, and whether they buy the newer models or stay with less capable, tried-and-true models will be one indicator of how quickly customers will accept the new satellites.

Standing Out From The Pack

In their marketing battles, each builder tries to differentiate itself from the pack. Boeing stands to gain much mileage from Hughes’ long-time industry leader role, with the most experience and highest number of sales almost every year. Hughes also makes a large investment in R&D, enabling it frequently to be the first to offer technology advances on its satellites.

Lockheed Martin positions itself as the only manufacturer that “does it all”–able to provide satellites, help with regulation and financing, and go to its sister company, International Launch Services (ILS), for launch. That said, the company is prohibited by government rules from having a relationship with ILS to get special treatment for launch services.

SS/L’s advantage is a “smaller” factory–although at 1.4 million square feet total, Loral’s factory is in fact very large–that allows very competitive pricing: “We are sized to break even,” says Klineberg.

Alcatel differentiates itself as the only builder owned by a large telecommunications company, giving it a unique, customer-oriented perspective on its products. “We don’t try to sell satellites to customers who might have their needs served with other means,” says de Bayser. “When we bring a solution to an operator, there has been a strong debate within Alcatel as to whether space makes sense compared to other solutions.”

Matra is promoting its high-quality design and its attention to fully complying with customer requirements. “The concept of Eurostar is ranked very high,” said Aubertin. Matra aims to cut its production costs by 20 percent, allowing it to become a low-priced bidder, as another differentiating factor.

Hughes

In early 2000, Boeing agreed to buy the world’s leading satellite maker from Hughes Electronics Co. for $3.75 billon.

After 40 years of building satellites, having delivered more than 100, Hughes recorded only five new satellite sales for 1999. That was down from 15 in 1998. The company will not disclose its maximum annual production rate, but said that at year end 1999, 37 satellites were in production, and 22 are schedulaed to launch in 2000.

Hughes’ 1999 backlog of unclassified satellites, commercial and government, was 38. The fact that the HS 702 had not yet flown in 1999 may have affected Hughes ability to sell the satellite. Nine 702s had been sold prior to the first launch in December for initial customer Panamsat. “When the 702’s performance is demonstrated, it will give us a competitive advantage over paper designs as the satellite market moves to more transponders, greater power and greater capacity,” Brinkley says.

Brinkley says Hughes’ strategy for winning new orders involves selling to repeat customers, who come back because they are pleased with on-orbit performance of the Hughes line. “Our focus is on establishing a heritage relationship with the customer. SES and Telesat are two examples. We’d like to reestablish a long-term relationship with Intelsat and Inmarsat, either as a prime contractor or preferred secondary supplier,” he says.

In 1999, export problems were followed by unidentified issues on the 702 and geostationary-mobile lines of satellites that drove up costs and delayed several satellites. These satellites, including ones belonging to Panamsat, which is majority-owned by Hughes, were delayed four to six months. Hughes also decided last June to stretch out its production time for some satellites because it was too fast and some tests were being sacrificed.

Hughes may be able to reduce the time and cost of production down even further with a concept called the “six-pack,” in which satellites are bought and built in blocks. The HS 601 could be built in 13 months if customers could be convinced to preproduce some standard parts and build the satellites six at a time, MichaelSmith, Hughes’ CEO said in December. “The problem is getting the customer to accept a standard satellite,” he explains.

Hughes also has a reputation for leading in the introduction of new technology, which can be a tough sell to customers who want the advances but are reluctant to accept unnecessary risk. The XIPS xenon ion propulsion system has flown for about two years on Hughes’ satellites, allowing customers to save weight in carrying fuel for stationkeeping in orbit. In the power area, Hughes is in the process of introducing “triple junction solar cells” to improve upon dual junction solar cells that have only been in use about two years. The newer cells will have 40 percent efficiency in converting sunlight to electricity, compared to 26 percent previously, which will bring more improvements in cost to the customer. Hughes claims it can double the cost effectiveness of its satellites every two to three years.

Hughes obtains much of its advanced technology from the Hughes Research Lab in Malibu, CA, owned by Hughes and Raytheon. This lab is a campus-like facility on the hills overlooking the Pacific Ocean where advances are made on semi-conductor chips, batteries and other essential spacecraft components. In late December 1999, Hughes researchers were working on a lithium ion thin film battery that would weigh one-half of current battery weight and have flexibility in shape so that it could fit in smaller corners inside the satellite. Other research was aimed at developing indium phosphide chips for higher-speed data rate transmission, and chips that have photonic and microwave circuits integrated together so they can use either fiber or satellite to carry a signal. Another chip project, funded by the Mayo Clinic, was merging four streams of 10 Gbps data into a single 40 Gbps stream, which would quadruple the fastest data rates now carried over the Internet infrastructure, according to Loi Nguyen, microelectronics lab fabrication operations and program manager.

Hughes managers are under pressure to constantly increase their internal efficiencies at the same time as they improve the product and keep customers happy. An ING Barings report last year said that since 1991, Hughes had doubled the revenue per employee from $150,000 to $300,000, and during the same years production time shrank from three to two years for many satellites.

Space Systems/Loral

Space Systems/Loral in Palo Alto, CA, won six new GEO contracts in 1999, company officials say. Loral can produce nine to ten satellites per year, according to Livanos. SS/L counts in its backlog several satellite wins that had not been announced publicly, such as the Assuresat satellite and a Telstar 8 with a 19 kW power level.

“We are running at 80 to 90 percent capacity,” Livanos says. SS/L proved in 1999 that it was up to juggling numerous customers who were ready for launch in a short window by delivering 11 satellites, including eight Globalstars, in a 30-day period for launch from three launch sites. In 2000, SS/L plans to launch its satellites at a rate of one per month.

In late December, SS/L had roughly 20 satellites in production and a backlog of 27, not including eight low-earth orbit Globalstars worth about $1.3 billion, that included several Asian satellites on which work had been stopped and one satellite, Chinasat 8, stuck in the factory by a lack of export approval. SS/L appeared to have been hit hardest by the Asian economic crisis and export control changes.

Delivery time can be as fast as 15 months, the time SS/L took to build an Orion satellite, Livanos says. The company carries a large inventory of parts, which helps reduce cycle time. SS/L buys long lead parts in advance, but satellites are not manufactured without a firm customer, who must be identified before frequency plans and antennas can be designed.

Collins expects new orders to roll in at a healthy clip in 2000, with prospects for contracts for Loral’s Cyberstar broadband satellites; Echostar for satellites to serve direct-to-home TV distribution; and an order for a new Brazilian satellite from Loral’s new company there. A Japanese customer is likely to buy a digital audio broadcast satellite, while other new orders should come from Japan’s Jsat, Optus of Australia and Japan’s MTSat (to replace the first one that was destroyed in a launch failure in fall 1999). Netsat28, a new broadband satellite licensee, also bought a satellite in early 2000; and Loral considers that a good prospect, although officials would not comment on the nature of the relationship between the two companies.

Since Loral acquired SS/L from Ford Aerospace in the early 1990s, manufacturing revenues have grown substantially, going from $600 million in 1993 to $1.4 billion in 1997.

While Loral executives claim that their satellites have the highest reliability record in the industry, the company continues to seek more redundancy and techniques to avoid failures in response to customer demands, Klineburg says. New satellites, for example, are being designed so they can function in orbit even if some systems fail.

Lockheed Martin

At Lockheed Martin Commercial Space Systems, the campaign to move up the ranks of manufacturers is being driven by aggressive bidding on contracts and the adoption of new management processes aimed at creating a team that knows how to please customers.

Parent company Lockheed Martin had a rocky year in 1999, with its stock price in decline much of the year and trouble inside several business units, including the satellite maker. Lockheed reported declines in sales in the first two quarters of 1999 of $130 million and $140 million in the sector that includes satellites and ballistic missiles, attributing a significant portion of the combined $270 million in decrease to “cost issues pertaining to technical issues on certain commercial satellite programs.”

After a four month review, an independent team found poor management oversight and quality control for the company’s space and missile businesses, according to a fall 1999 report to Lockheed. Specifically, the panel said the company had problems in its accountability, in managing its suppliers and subcontractors, and in emphasizing cost control over quality. Much of the trouble was found in the launch vehicle programs, particularly the Titan 4 used by the military.

But for LMCSS, 1999 also marked achievements, like the launch of the LMI 1 satellite for Lockheed Martin Intersputnik, and a big win from GE Americom in February 1999 to build four new A2100s.

In Sunnyvale, just south of San Francisco, Lockheed Martin operates an enormous satellite factory, opened in 1996, with a class 100 clean room the size of two football fields that is said to be the largest in the world. LMCSS’s work on commercial communications satellites combines heritage from two areas. Lockheed Martin has decades of experience building classified satellites for the U.S. government, enabling it to claim a role in producing more than 850 satellites total. On the commercial side, the company acquired the satellite units from the former RCA Astro and GE, moving those operations to California in the early 1990s.

In December, nearly a half dozen satellites were in the final stages of assembly at Lockheed’s plant, and company officials said they had 14 satellites in some stage of manufacturing, with customers including Astrolink, New Skies, Space Communications Corp., Japan Satellite Systems Inc., NTT and GE.

Production time for a LMCSS A2100 satellite is down to a minimum of 18 months. LMCSS is capable of producing eight satellites per year in its factory with no further investment, and 12 if more investment is made. In 1999, LMCSS delivered six geostationary satellites, and in 2000, six more are scheduled for delivery. The factory is running at 75 percent capacity.

The company’s continuous improvement effort involves training its team in management techniques like the Kaizen process, which fundamentally changes a process by tearing it down, then building it up with a focus on the objective, and eliminating all steps that do not add value to the objective. John Zedro, LMCSS’s director of strategic initiatives, says production time for subsystems and components, and the number of defects, have dropped as a result. The philosophy goes: “If it doesn’t add value to the customer, don’t do it,” says Zedro.

In the technology area, LMCSS is developing dual and triple junction solar cells that could increase efficiency from 13 to 15 percent to the 25 to 30 percent range. In the area of active array antennas, LMCSS is developing a design where antenna elements can be built into tiles that can be fit onto a satellite. Inside the tiles, these elements radiate a signal from a beam and amplify the signal, according to David Bair, LMCSS director of advanced concepts. The technology may be ready for bidding to customers in 2001.

Alcatel

Alcatel Space had a slow sales year in 1999, winning only one new geostationary satellite contract, the Hot Bird 6 satellite for Eutelsat. Alcatel Space’s de Bayser says three “internal” contracts, two for Europe*Star and one for Eurasisat, also are on the books as 1999 orders. Another payload subcontract for Nilesat, from prime contractor Matra Marconi Space, also was booked in 1999.

Alcatel serves as prime contractor and supplies payload equipment for communications satellites from its manufacturing facilities at Cannes and Toulouse, located in the southeast and southwest of France respectively. The company was created with the July 1998 merger of Alcatel, Aerospatiale and Thomson CSF.

At end 1999, Alcatel Space had six communications satellites in production, plus payloads for five more. Alcatel Space is capable of delivering eight satellites a year from its factory, according to de Bayser. To profit, Alcatel Space must run at 50 percent capacity, he says, adding that Alcatel Spacedoes a lot of work on large communications payloads and other equipment for other prime contractors like Loral, which helps to fill up the factory. Alcatel Space does not disclose its current backlog.

Alcatel Space’s strategy is two-fold, with some wins coming as the result of investment in a customer and others in the open market from “improving on a permanent basis and delivering more powerful satellites,” with the Spacebus 4000 being the key element.

The year 2000 promises to be a much stronger year if Alcatel can get Skybridge under contract. The Skybridge system of 80 satellites, plus supporting infrastructure, will cost $4.8 billion, and will require a strong surge in component manufacturing capabilities to meet its schedule. This will also be a busy launch year for Alcatel Space, with 10 satellites for which Alcatel Spaceserved as prime contractor scheduled for launch, plus eight more for which the company provided equipment.

Alcatel Space’s advantages include its proven track record in producing LEO satellites for Globalstar, which required a rate of one per week, a pace similar to that expected for Skybridge, and a demonstrated willingness to invest in and help finance its customers. Alcatel Space has a long list of investments in its manufacturing clients that includes Cyberstar, Globalstar, Eurasiasat, Europe*Star and Rascom.

Alcatel Space says Spacebus has never had a mission failure and has an overall 96 percent reliability record for the platform, which it believes is the best available.

Alcatel Space also is willing to buy long lead parts for satellites that haven’t yet been ordered. “There is the question of taking financial risk before we get an order, and we are doing it more or less,” de Bayser says. Gambling that a customer who has bought one satellite will need another is part of the strategy, with the fallback position being if the original prospect does not want the satellite, another buyer can usually be found. With the early ordering of some hardware, a satellite can be delivered as fast as 12 to 14 months, de Bayser says.

Matra

Matra Marconi Space (MMS) is transitioning from its tried-and-true line of Eurostar 2000 satellites to the more powerful Eurostar 3000s, and while the company began bidding the 3000s almost exclusively in 1999, the year did not produce much in the way of new orders. Only one geostationary telecommunications contract, for Nilesat, was signed for last year. “It was a transition year,” says Aubertin, who expects 2000 to bring a larger number of geostationary orders.

“We base our strategy on full compliance to the customer requirements, and on the high reliability and quality of the satellite. For the Eurostar 2000 family, no mission has been jeopardized” due to performance, Aubertin says.

Matra also is refusing to compress production times to the bare minimum. “With the failures in 1997-99, customers are putting emphasis on getting a good satellite rather than a medium-quality satellite two months sooner,” Aubertin says. Matra has delivered satellites in 18 and 22 months, but believes the standard production time for a 3000 series satellite will be 27 to 32 months, which Aubertin says is acceptable to customers.

Despite the low sales, the MMS factory is close to its annual production capacity of eight satellites per year due to its practice of balancing the workload by building remote sensing and telecom satellites in the same plant. Matra, with two sites in France and four in the United Kingdom for satellite manufacturing, was formed from the satellite units of Matra Espace, Marconi Space Systems and British Aerospace Space Systems. This year, the venture will take the name Astrium, when MMS is joined by the Daimler Chrylser space activities. With their forces combined, the new company should have annual revenues of $2.3 billion and an 8,000-person workforce. The MMS satellite manufacturing revenues in 1998 came to about $1 billion.

In 2000, Matra will launch seven telecom satellites for customers that include Worldspace, Nilesat, Eutelsat and SES Astra. In 1999, MMS began a program to reduce manufacturing costs, which should cut its base prices by 20 percent, enabling the company to begin positioning itself in the market as a low-cost supplier. Part of the effort is devoted to parts suppliers, who are being converted into “partners” with a greater stake in each project, and also getting bulk orders for parts for the Eurostar 3000 series to lower costs.

Within the next five years, the 2000 series production will taper off to one to two per year, built for regional operators or to serve bigger operators who need fast, auxiliary capacity. But most of Matra’s efforts are aimed at establishing the 3000 line. For WEST (Wideband Europe Satellite Telecommunications), Matra is working with satellite operators to whom it is propoaing its WEST broadband solution based on geostationary satellites, Aubertin says.

More Orders in 2000

In 2000, several factors could prop up the satellite manufacturing business, including bulk orders for low earth orbit constellations, the launch of new technology on mobile geostationary satellites for the first time, and the restart of the Asian market. Executives say 2000 certainly will be better than 1999 was for new orders. One thing is for sure: The battles among satellite builders for the handful of openly competed contracts will be hard fought, and the winners will be the largest players with the most competitive strategies.

Theresa Foley is Via Satellite’s senior contributing editor.

---