For 25 years, the United States has poured billions into a “virtual wall” of surveillance towers along the southern border, with officials presenting the network as a way to detect, apprehend and ultimately save people crossing the frontier.¹

But an MIT Technology Review investigation, produced with the Times of San Diego, argues that the promise has repeatedly collapsed in the desert. Reporters documented more than 1,050 people who died near surveillance towers—an acknowledged undercount, because Customs and Border Protection does not publish a complete map of the network and the project examined nearly 600 towers, not all 800 the government says are operating.²

The reporting found people moving through tower-covered areas without being reached or apprehended, including near newly installed AI-powered systems intended to spot people automatically. In some cases, bodies went unnoticed for weeks or months. The failures were not framed as a single technical glitch: the investigation cited broken towers, algorithms that did not detect people, and agents who did not respond to alerts.²

The gaps extend beyond detection. Border Patrol’s system for recording migrant deaths was not governed by national entry rules until 2021, and the Government Accountability Office subsequently found that many cases—particularly those not discovered by Border Patrol—were still missing. The agency’s Missing Migrant Program has improved its data collection, the investigation says, but lacked a plan to measure whether it was actually increasing rescues or reducing deaths.²

Now the stakes are rising. The U.S. is set to spend $1 billion to triple the virtual wall by 2034, even as an internal assessment of tower effectiveness remains out of public view. MIT Technology Review’s prescription is less about adding cameras than forcing accountability: audit deaths near towers, repair death tracking, log technology-assisted apprehensions, and automatically investigate whether every nearby death reflects a surveillance failure.²