The automated train control system is an interesting engineering/financial problem. I have a electrical engineer acquaintance who installs these here in the US. It is apparently a long involved process. He spends about a year doing a segment of track. I haven't talked with him about the details (what is involved in installation, length of a segment, number of hardware pieces, amount of customization required etc.).
The interesting question comes from taking this testimony at face value. There are two choices parallel and serial. Creating a large industry (workforce, infrastructure and so on) that will be largely shut down when installation is complete as maintenance can be expected to be a much smaller job, or spreading the installation over time. The first choice minimizes wrecks and the second minimizes investment and economic dislocation of those in the industry. While you can also bring in things like invention to speed the installation, the same fundamental problem will exist. And obviously the train system in the US has opted for the second choice. Good for this friend as it seems likely the job will last the rest of his life.