It is one of the first things they teach on physical security that barriers are incapable of keeping people out; that thinking the existence of a wall or door will keep someone out is the absolute worst mistake you can make. There is no such thing as an effective barrier.
Instead, doors, fences, and locks are delay devices, meant to keep someone busy while the the alarms bring responders. In any real physical security scenario, it is really the responders (and the threat of them) that keep people away, not the walls.
Further, when there IS a wall, generally all costs for security and alerting are built into the wall, offering a fantastic opportunity for attackers to plan distractions, costly vandalism, and to circumvent the easily probed security. You k ow how expensive it is to hire a welder to work in the desert, and to haul repair materials out like that? It's certainly not cheap. And then you need an HSI expert to repair and test the security.
But the desert doesn't need another delay mechanism. The desert there itself is the delay mechanism, as it takes days to cross, and there's nowhere to hide.
Instead, we just need an effective alarm. Instead of a wall or barrier, what we need is drones: cheap, efficient cameras capable of quickly being replaced and covering large areas. Put a motion sensor and anomaly detection on the drone platform, have it call out a second drone for continued monitoring, and follow the incursion while alerting the actual border patrol to send out a truck.
This would be far more effective. If a drone goes down or gets shot down, it would be trivial to detect and send replacements and response. All the drones need is a platform for recharging: put them on the border patrol vehicles.
We don't need another delay mechanism; we already have one. We just need an effective alarm.