The concept of having a heater in a shipping container that initially generates 15 kW of heat and only slowly (years) will have a higher power capacity is commercially not attractive.
The construction is dimentioned such that it can deliver 500 kW, so why not control the number of fusions in such way that it is able to generate 500 kW all the time.
I would be much more interested in technology that converts the annihilation energy rather than fusion energy.