On the night of February 15, 1996, a rocket lifted off, hung on a pillar of fire, swayed — and disappeared behind a hill. A moment later, an explosion lit up the sky over Sichuan. This is the story of how a technical failure in an inertial platform turned into the loudest American-Chinese technology scandal of the post-Cold War era.
Long March 3B — the flagship three-stage rocket of China's Chang Zheng program — launched for the first time from Xichang cosmodrome at 03:01 local time (February 14 UTC). On board was Intelsat 708, an American telecommunications satellite insured for $204.7 million, assembled by Space Systems/Loral. The rocket was supposed to place it into geostationary orbit — a standard commercial mission under the growing US-China cooperation in the space launch market.
Two seconds after liftoff, the trajectory began to curve. The rocket lost vertical, tilted left, as if an invisible hand had shoved it sideways. Telemetry showed normal thrust from the main engines, but the course was already beyond correction. 22 seconds after launch, Long March 3B crashed into a hill 1.8 kilometers from the launch complex and exploded, turning the satellite into a cloud of metal fragments.
Official Chinese data recorded 6 dead and 57 injured. Western journalists who arrived the next day saw destroyed houses, craters from debris, and closed zones where no one was allowed. In subsequent years, estimates of "dozens to hundreds dead" began appearing in books and articles, but documentary confirmation of these figures never surfaced. Names of victims — Qian Zhiying, Yang Linzhen — appeared in Chinese official materials years later, as if the state were slowly opening the archives of the disaster one page at a time.
The initial Chinese investigation blamed the accident on "inertial platform failure" — a formulation so vague it could mean anything: from a software bug to mechanical damage. But the real story began when the Americans joined the investigation.
The Independent Review Committee (IRC), created with the participation of American specialists from Hughes and Loral, conducted its own technical investigation. The Chinese side provided debris from the control system, telemetry, and pre-launch test records. The Americans weren't just looking for a cause — they were looking for a mechanism.
The committee established: the disaster occurred due to defective soldering of gold-aluminum connections in the electrical circuits of the guidance system servo. These microscopic contacts were responsible for transmitting commands from the inertial platform to the engine steering nozzles. One of the connections turned out to be fragile — either due to poor quality material or violation of temperature conditions during assembly. Under launch vibration, it failed, breaking the command chain.
In the Western press, the technical side of the accident quickly became overgrown with rumors: some attributed the rocket's hypersensitivity to wind shear — a version American companies had been circulating for years after the Apstar 2 accident two years earlier — others spoke of defective fasteners allegedly causing a chain reaction of failures. Nice stories, but the IRC report contained neither. The problem was in microelectronics, not mechanics. Rumors were born from simplifications and retellings, but stuck in popular culture as a symbol of "Chinese negligence."
The actual technical picture looked different. The Long March 3B guidance system used a platform with gyroscopes that constantly tracked the rocket's position in space. When it deviated from the calculated trajectory, servos were supposed to rotate the engine combustion chambers, creating a control moment. When one of the drives lost connection with the platform, the rocket became a free-flying projectile with working engines — controlling it became impossible.
IRC completed its work within several months after the disaster, but its report never became a public document. The Chinese side classified the results, the American side didn't demand full disclosure. Instead, another story began: an investigation concerning not what went wrong with the rocket, but what could have gone wrong with secrets.
The US Congress wasn't interested in defective soldering. Congress was interested in whether Hughes and Loral had transferred information to China capable of improving the reliability of Chinese ballistic missiles. The logic was simple: if American engineers participated in analyzing the causes of the accident, then they discussed the structure of the guidance system with their Chinese counterparts. And a commercial rocket's guidance system differs little from an intercontinental ballistic missile's guidance system.
A special commission created in 1998 under the leadership of Congressman Christopher Cox completed a classified report by January 1999, and in May 1999 a declassified version came out — three volumes totaling about eight hundred seventy pages. The main conclusion: China systematically obtained classified American technologies, using commercial space launches as cover for industrial espionage. The Intelsat 708 disaster became a key episode in the accusations.
The report claimed that engineers from Hughes and Loral, participating in IRC work, provided the Chinese side with detailed technical recommendations for improving control systems. These recommendations weren't military secrets in the strict sense — they concerned eliminating known design defects. But from the standpoint of American export controls, any information capable of increasing missile guidance accuracy automatically fell under a ban on transfer to a potential adversary.
The companies denied violations. Loral insisted that all discussions were conducted within the framework of State Department-authorized cooperation, and technical documents were transferred with the consent of American authorities. But the Cox Commission found correspondence in which engineers discussed details of gyroscopic platform construction and power supply circuits — topics that formally required separate export authorization.
In 2002, the investigation ended with a settlement agreement: Space Systems/Loral paid $20 million — $14 million in fines to the State Department and $6 million for expanding export control compliance programs. This was the largest fine in the history of American space export controls — and a signal to the entire industry.
In March 1996 — already after the Intelsat 708 disaster — administrative control over commercial satellite exports transferred from the State Department to the Department of Commerce: they were listed in the Commerce Control List (CCL), and obtaining a license was bureaucratic but not prohibitive. China was a large market: low launch prices, growing demand for geostationary slots for telecommunications, state guarantees of contracts.
By 1999, amendments to the National Defense Authorization Act adopted by Congress moved all commercial communications satellites to the category of military products — the United States Munitions List (USML). Now export control was carried out by the State Department under International Traffic in Arms Regulations (ITAR) rules, developed for tanks and missiles, not telecommunications equipment.
Practically, this meant the end of American-Chinese commercial launches. Getting ITAR permission to launch a satellite with American components on a Chinese rocket became nearly impossible — the procedure required proof of no risk of technology transfer, which during joint work on accident investigations was unachievable. Hughes, Loral, Boeing — all major satellite manufacturers lost the Chinese market in one legislative cycle.
European manufacturers gained a competitive advantage. Airbus Defence and Space (then Astrium), Thales Alenia Space — companies not bound by American export controls — continued working with China. But they too faced restrictions: many critical satellite components (processors, sensors, radio transmitters) were manufactured in the US and fell under ITAR, even if the final satellite was assembled in Europe.
The side effect turned out to be larger than direct bans. American companies began losing market share globally: clients from countries not in the US alliance preferred ordering satellites without American components to avoid ITAR bureaucracy. By the mid-2000s, the US share of the global commercial satellite market fell from approximately 83% to around 50%. Lawmakers were protecting national security but undermining national industry.
China didn't become a space superpower because of sanctions — that's a myth convenient to both sides. Beijing did indeed begin an accelerated import substitution program after 1999: development of its own navigation systems (Beidou), domestic satellite components, new generations of Long March launch vehicles. But these programs were conceived long before the Intelsat 708 disaster — the first Beidou plans date to the mid-1980s, and the concept of a new-generation heavy rocket was being worked out back in the late 1980s as part of Program 863, though the Long March 5 project only received state approval in 2004.
Sanctions accelerated the process but didn't create it. China's space program was always a product of state strategy, multi-year investments, and industrial policy. By the 2020s, China had put its own modular station Tiangong into orbit, landed a probe on the far side of the Moon, launched a Mars rover. These achievements are the result of decades of work, not a reaction to loss of access to American technologies.
For the US, the consequences turned out to be ambiguous. ITAR did indeed reduce the risk of leaking militarily significant technologies through commercial projects, but the price was high: American space industry lost markets, competitiveness, and pace of innovation. In the 2010s, Congress began discussing export control reform, trying to find a balance between security and economics, but the ITAR structure remained almost untouched.
Western estimates of the number of disaster victims remained unconfirmed. The Chinese official figure — 6 dead — appears in all state documents, but details (exactly where people were located, how they died, why destruction was localized) were never disclosed. Journalists who visited Xichang in subsequent years found traces of destruction that didn't correspond to the scale of official data. The truth remained somewhere between six and "hundreds" — in a zone where documents end and eyewitness testimony and closed archives begin.
Twenty-two seconds of flight turned into twenty years of consequences. Defective soldering in a microchip launched a chain reaction: technical failure escalated into political scandal, scandal into legislative changes, changes into restructuring of the global space market. The rocket fell on a hill in Sichuan, but the shockwave reached Washington, Brussels, and back to Beijing, changing the rules of the game for all participants in the space race for decades to come.