PARIS — Satellite manufacturers say they have been able to introduce new technology into their commercial products in recent years despite the natural conservatism of their customers and of insurance underwriters, who despite recent profitable years continue to live in fear of a repeat of the string of satellite failures that occurred starting in the late 1990s.

The result, the manufacturers say, is higher-performing spacecraft whose failure rate has remained low in terms of outright in-orbit failures that have resulted in insurance claims.

The new technologies include lithium-ion batteries, which are becoming standard on many spacecraft; gallium arsenide solar cells; electric propulsion for in-orbit satellite station-keeping; higher-precision antenna-pointing systems; and ever-larger unfurlable antennas for mobile applications. The latter innovation has advanced to where the 9-meter-diameter antennas on Inmarsat satellites now seem quaint compared with the 18-meter mesh antenna deployed on the TerreStar Networks satellite built by Space Systems/Loral and launched in September 2009.

The TerreStar antenna, built by Harris Corp. of Melbourne, Fla., is about to be relegated to second-place status by the two 22-meter antennas Harris is building for a pair of Boeing-built SkyTerra satellites to be launched in the next 12 months.

The progression to ever-larger antennas, which permit customers to use smaller, lower-power communications equipment, has not been without glitches: The 12-meter-diameter Harris antenna launched aboard the Eutelsat W2A satellite in May malfunctioned for reasons not yet clearly explained, forcing the customer, Solaris Mobile of Ireland, to scale back its business development plans.

The satellite industry has long been one in which production engineers push for improvements that cannot always be tested in orbit beforehand, and satellite fleet operators and their insurance underwriters resist being the first, or even the second, customer to try out the new technology.

“Customers want to see at least three years in orbit without anomalies before adopting new technology,” John Celli, president of Palo Alto, Calif.-based Space Systems/Loral, said during a March 15 panel discussion at the Satellite 2010 conference in Washington.

But the hesitation of many satellite operators to be technology pioneers does not necessarily apply to those starting new businesses whose success may depend in part on hardware that has not been fully proved in orbit. That is the case for several mobile satellite service providers, among others.

Chris Hoeber, senior vice president at Loral, noted that the company has launched three satellites with large unfurlable antennas — XM-5 for Sirius XM Radio, TerreStar-1 and ICO G1 for DBSD North America — since 2008, all without a problem.

The amount of new technology that an operator can accept depends in large part on the satellite’s intended mission, and the operator’s risk and schedule tolerance.

Martin Gee, chief technology officer for Yahsat of the United Arab Emirates, a startup operator with two large spacecraft on order, said Yahsat has inserted flexible dates into its customer contracts to be able to adjust to any delays in its satellites’ construction at Astrium Satellites and Thales Alenia Space, both of Europe. That reduces pressure to rush production, he said.

In an address to the World Space Risk Forum in Dubai, United Arab Emirates, March 1, Gee, who was formerly with Paris-based Eutelsat, said operators are occasionally willing to adopt new technology if there is a backup. For example, the early failure of ion-electric propulsion on several commercial satellites has slowed that technology’s full introduction.

Gee said operators still want to have at least some conventional chemical propellant available in the event the electric-propulsion system fails. The recent news that Satmex of Mexico lost the second of two xenon-ion propulsion units on its Satmex 5 satellite and is relying on a backup chemical-propellant system — providing nearly three years of life to a satellite that otherwise would be out of service — will not hasten operators’ willingness to rely solely on electric propulsion.

Philippe Pham, head of industrialization at Astrium Satellites, said market demand for higher-power satellites will move manufacturers to introduce technologies that make it easier to deliver that power to the payload without exceeding the weight and volume ceilings imposed by the current fleet of commercial launch vehicles.

At the space risk conference, Pham said Astrium expects that half of the geostationary-orbit commercial satellites ordered in the coming five years will require more than 8 kilowatts of end-of-life power to their payloads.

Guy Perez, vice president for telecommunications at Thales Alenia Space, agreed that customers are asking for features, including high-precision antenna pointing, multiple spot beams and even four-axis steerable antennas — three axes plus a zoom function — that will automatically force new technology development.

Astrium was among the first Western satellite builders to introduce plasma-electric propulsion into its product line, succeeding in part by convincing customers that the weight savings over chemical propellant could be put to use to improve the satellites’ payload.

Astrium and Thales Alenia Space, with substantial financial support from the French and European space agencies, are jointly developing a satellite platform called AlphaBus that will match or exceed Loral’s highest-power LS 1300 model and deliver 25 kilowatts to the payload.

But beyond a first, 18-kilowatt model ordered by London-based Inmarsat, AlphaBus so far has not found its commercial footing.

Peter B. de Selding was the Paris bureau chief for SpaceNews.