"Ansar Allah’s Missiles Have Learned to Hit.
Someone Taught Them"
by Larry C. Johnson
"For most of the Saudi war in Yemen, Ansar Allah’s ballistic missiles were more political than military. They were noisy and frightening, but rarely decisive. In 2017 the press could fairly report that the group’s missiles were often lightened by reducing their payloads and seldom hit what they were aimed at. Saudi air defenses picked off many of them, and the rest fell in empty desert.
That is no longer the picture. Since the confrontation reignited this summer, five Saudi airports have come under Houthi attack since October 5, with 20 dead and hundreds wounded. On October 10, AFP photographed King Khalid International Airport in Riyadh at the moment it was struck. When the Houthis claimed hits on Aramco facilities, the coalition called the claims misleading, but independent journalists reported fires and smoke over the Riyadh refinery. These are not area weapons landing somewhere near a city. They are hits on specific facilities at ranges of 1,000 kilometers or more, against one of the densest air-defense networks in the region.
I spent a career around targeting and military planning, and I know what that pattern means. Accuracy at long range doesn’t come from a better rocket motor. A Scud-derived missile flying on inertial guidance alone accumulates error throughout its flight, and its reentry vehicle adds more. Getting from a miss of several hundred meters to tens of meters requires correcting the trajectory late in flight, through satellite-navigation updates, a maneuvering reentry vehicle, or a terminal seeker. Whatever the hardware, it also requires precise target coordinates. Neither of those is something Ansar Allah can manufacture in the mountains of Saada. The missile bodies are Iranian lineage. That is not seriously disputed. But Iran’s own precision gains point to their source, and here there are two candidates, not one.
The first is Russia. Since Iran’s 2020 strike on Ain al-Asad, there have been persistent reports that Fateh-family missiles use GLONASS, Russia’s satellite navigation system. For Iran to use the encrypted, jam-resistant military GLONASS signal, Russia would have to hand over the cryptographic keys.
The second is China, and I would not discount it. China’s BeiDou constellation is larger than GPS, and its restricted, authorized-user service is built to resist jamming and spoofing. Researchers who follow the China-Iran relationship point to a reported 2015 agreement to integrate BeiDou into Iran’s military infrastructure, and to reports that Beijing opened its encrypted military signal to Tehran after the 2021 strategic partnership. By some accounts Iran completed its move off American GPS and onto BeiDou in June 2025, right after the 12-day war. Former French intelligence chief Alain Juillet has argued that the improvement in Iranian missile accuracy since then is best explained by BeiDou access. Tehran has never confirmed any of this. It says only that it uses every available source and depends on none.
I suspect the honest answer is that Iran uses both, and that whichever constellation its missiles rely on, the same logic follows. The open civil signal is exactly what Saudi and American electronic warfare would jam first. A missile that keeps hitting through that environment is using a protected military signal, and only Moscow or Beijing can grant that. If Iran has it, there is every reason to believe the missiles it ships to Yemen carry it too.
The targeting layer points the same way. In 2024 the Wall Street Journal reported that Russia was passing satellite data to the Houthis through the IRGC in Yemen. Washington has also publicly accused a Chinese commercial satellite firm of supplying imagery that supported Houthi attacks. Precise coordinates are half of precision strike. A movement with no satellites of its own needs someone else’s eyes.
“They just use Google Earth”: Some commentators wave this away. Anyone can pull coordinates off Google Earth, they argue, so Ansar Allah needs no outside help. Type in the coordinates of a runway or a refinery and fire. That argument confuses knowing where a target is with being able to hit it.
Accuracy has two separate error budgets, and military targeting doctrine treats them separately. The first is target location error: how well you know where the aimpoint is. The second is weapon error: how closely the missile can actually fly to a given coordinate. The two combine, and the larger one dominates. Free commercial imagery has improved the first. For a large, fixed, well-imaged facility like a Saudi airport, Google Earth coordinates are often good to within a few meters horizontally. That’s a real change from the 1990s, and I won’t pretend otherwise.
But a coordinate is just a number. It does nothing for a missile that can’t steer to it. A Scud-derived missile flying on inertial guidance, with a reentry vehicle that wobbles as it comes down, misses by hundreds of meters no matter how perfect the coordinate is. A 5-meter coordinate fed into a 500-meter weapon gives you a 500-meter weapon. The only way to turn good coordinates into good hits is a guidance system that can correct its trajectory late in flight: satellite-navigation updates, a maneuvering reentry vehicle, or a terminal seeker. The Google Earth argument doesn’t explain away the need for that technology. It assumes the technology is already there.
There’s a second problem. Satellite-navigation guidance is only as good as the signal it receives, and the civil GPS signal is the first thing Saudi and American electronic warfare jams in wartime. A missile that keeps finding its target in a contested electromagnetic environment is not running on a free, unencrypted signal. Access to a protected military signal, whether GLONASS or BeiDou, can only come from the state that owns the constellation.
Finally, commercial imagery is a snapshot, often weeks or months old. It can’t tell you which hangar holds aircraft tonight or which part of a refinery is running. Hitting high-value aimpoints rather than just facilities requires current intelligence, which is exactly what the Russian satellite data reported in 2024 would provide.
So yes, Google Earth is part of the story. It helps explain why targeting fixed infrastructure has become easier for every non-state actor in the world. It does not explain how a Yemeni movement’s missiles started arriving where they were aimed. That requires a guidance package and a navigation signal, and Ansar Allah makes neither.
The chain of supply: So the chain is not hard to reconstruct. Iran provides the missiles, the training, and the IRGC advisers on the ground. Russia, China, or both provide what the United States cannot switch off: a protected military navigation signal, GLONASS or BeiDou. Russia has already been reported providing targeting data, and China supplies commercial imagery. Ansar Allah provides the launchers and the will.
Why would Moscow or Beijing take that risk? Look at the map. The United States has two carrier strike groups and an amphibious ready group in the Arabian Sea for its war with Iran. Every Saudi air-defense interceptor fired at a Yemeni missile is one not available for that fight, and every disrupted Saudi airport and refinery raises the political cost of Riyadh’s alignment with Washington. For Russia this is the same logic as the Western weapons and targeting support that has flowed to Ukraine. For China it is a low-cost way to keep American naval power tied down far from the Western Pacific. Neither needs to fire a shot or put its name on a missile.
Washington and Riyadh should stop treating the Houthi missile threat as an Iranian problem alone. Sanctions and interdiction at the Red Sea ports address the missile bodies. They do nothing about signals coming from orbit or imagery bought through a front company. Until the guidance and targeting layer is confronted directly, through diplomacy with Moscow and Beijing or through serious electronic and space countermeasures, Saudi airports will keep making the evening news."
Here is my latest Transition Protocol:
