Does it mean that at 0ns there is one path, 10ns second path, ... and approximately at 130 ns, there is a power path?
I've never heard of the term power path before applied to channel impulse responses. If my guess is correct, that figure depicts the CIR you get at different distances from a point, which can be the source or anywhere else. All the spikes you see at Xns are propagation paths. That is, you will receive a copy of the signal delayed Xns for every single peak.
It is not very clear from the figure but the main energy arrival seems to be a 100ns.
If you plan to use the CIR for comms simulations you will need to normalize the arrivals using the system bandwidth.
In the multi-path scenario we have occurrence energy from (typically) all directions, and a channel could be a Rayleigh fading channel; no LOS component or Rician fading channel; a strong LOS component is present.
The figure you referred to contains a dominant component with a delay spread of 130 ns and with responses from other paths occurring with different delay spreads. Also, the key to the figure mentions an RMS delay spread of 65.9 ns, which can be justified by noticing the higher power distribution between 0 to 100 ns delay spread.
Take a look at the picture below to better understand the CIR for a multipath scenario. In this case, the CIRs from multiple antennas will be combined to form the CIR of the antenna array.