Saudi Arabia’s bet on American military AI, the Maven Smart System, is now at the center of its defense against the Houthis, and the deal amounts to a costly capability transfer rather than a conventional arms purchase. The kingdom is buying speed: faster targeting, faster strikes, and a shrinking of intelligence lag that once favored its adversaries.
The arrangement took shape under pressure. US President Donald Trump reportedly declined when Saudi ruler Mohammed bin Salman asked him to take direct military action against the rebel group, according to accounts of the exchange. What Washington is supplying instead is intelligence, including an AI-driven targeting tool, plus personnel. A September report by US broadcaster CNN, citing several unnamed American officials, put the number of US soldiers stationed in Saudi Arabia for a new joint force command at between 100 and 200.
Those personnel are helping the Saudis “use a targeting tool that allows military officers to see in real-time what is happening on the battlefield,” CNN reported. Officials say the system deployed in the kingdom resembles the Maven Smart System, or MSS, which the US military itself has used to seek targets in Iran, Iraq and Syria.
The economics of the technology favor scale. MSS emerged from Project Maven, the Pentagon’s flagship military AI program launched in 2017, and analysts at the Center for Strategic and International Studies in Washington described it in a June paper as an AI-enabled digital platform that “aggregates, organizes and visualizes vast streams of intelligence, surveillance and reconnaissance data.” The software detects patterns in inputs such as live satellite feeds and drone cameras, flags potential targets like rocket launchers, tanks or suspicious vehicle convoys, and even calculates the best method of striking them. Work that once required human analysts to comb through hours of footage, labor CSIS researchers called difficult and often nearly impossible, is now compressed into minutes. Tech commentator Bilawal Sidhu, a former Google product manager, likened the setup to a “Google Earth for war.”
The return on that automation, measured in strike volume, has already been demonstrated. When the US military used MSS in its late February attack on Iran, the system facilitated hits on more than 1,000 targets within the first 24 hours of the war, compared with roughly 100 targets a day previously.
Meanwhile, the strategic balance in Yemen has moved against Riyadh. The civil war had been comparatively quiet after an informal ceasefire negotiated in 2022, with the Houthis holding large parts of northern and western Yemen since 2015 and the internationally recognized government, backed by Saudi Arabia and the United Arab Emirates, controlling the south and east. But the Houthi campaign that began in September has extended their grip along Yemen’s coast and brought renewed attacks on Saudi infrastructure, including strikes on the country’s main airport and oil facilities, assets central to the kingdom’s revenue base. Military campaigns by Saudi Arabia, the United States and others have so far failed to significantly weaken the group.
Farea al-Muslimi, a Yemen-focused research fellow at UK think tank Chatham House, told DW the Houthis take a “zero-sum approach,” rooted in an ideology that treats poverty and hardship as a divine test. “They don’t care if the Israelis bomb the airport or the US bombs a hospital,” she said. “It actually serves them very well because their ideology is all about martyrdom and sacrifice.”
By contrast with the theater it was built for, MSS faces hard technical limits in Yemen that could erode the investment’s value. The Houthis are highly mobile, using truck-mounted ballistic missiles and small drone swarms from hills and canyons in rugged terrain, and fighters may be long gone before a strike lands even with AI assistance. MSS works best in relatively unpopulated areas, and desert terrain changes, dust storms, high wind, snowfall, dense foliage and deliberately planted decoys can all degrade its performance. Cheap decoys, in effect, impose a cost on an expensive system.
The ledger on human costs is equally contested. A human is supposed to verify every AI-flagged target, yet among the sites MSS targeted in Iran on the first day of the war was a school, where an American missile killed almost 200 people, most of them children aged between seven and 12.
Israel, the region’s most advanced military AI operator, has used a comparable system called the Gospel, or Habsora, in Gaza. Analysts Frederic Wehrey and Andrew Bonney wrote for War on the Rocks that while AI there enhanced commanders’ situational awareness and strike speed, those benefits “have been overshadowed by mounting civilian deaths reportedly resulting from this technology’s safeguards being loosened or eliminated.”
Antony Loewenstein, an independent journalist based in Australia and author of “The Palestine Laboratory: How Israel Exports The Technology of Occupation Around the World,” doubts the technology will deliver returns against the Houthis. “I don’t think there’s much evidence at all that using an AI tool is going to make much difference on the ground, except for causing huge amounts of civilian casualties,” he told DW. “And I think what the US and the Saudis will find is that the Houthis have an incredibly high tolerance level for destruction.”
Whether a targeting platform designed for state-on-state warfare can generate returns against an adversary that prices destruction into its strategy remains the open question hanging over Riyadh’s newest defense investment.