> The other angle to consider is that tracks very likely won't be singular. The tubes are light by infrastructure standards, you could pack two for each direction onto a route without dramatically increasing pylon costs.

This observation fails to take into account real-world requirements of building this sort of infrastructure. Pylon costs (queue the starcraft references) are irrelevant. The relevant bit is building all the tunnels, bridges, and earthworks required to get a smooth-enough train route that is able to safely support the track, provide a safe ride at Hyperloop's projected speeds, provide a conformable ride, and do it all in cost-effective an easy to maintain technology.

In high-speed railway, track defects are measured in wavelengths of fractions of a milimeter in amplitude and periods of 6 to 10 to 50 meters. If you increase the speed, either the limit for the wavelength period increases or the amplitude limits need to be further reduced. Setting up a 200-meter track segment with a sub-milimeter tolerance limit isn't cheap.