Your analysis is absolutely implausible and not more than wishful.
Who will cumulate Pd? In what CF philesociety?
The palladium would accumulate in machines, just as lead accumulates in batteries, and palladium accumulates in catalytic converters today. Nearly 100% of the lead is recycled, but unfortunately, much of the palladium is lost when catalytic converters are used, and more is lost when they are recycled.
A cold fusion cell resembles a sealed battery, so it is likely nearly all of the pallidium in it can be recycled when the cell is scrapped.
Look the first MW unit will need 10kg Pd that
today 239000 $, add to this the price deuterium and find an investor for this disaster.
Your arithmetic is wrong.
I estimate that palladium can produce ~200 W/g, so you would need 5 kg. This costs $119,000 at today's prices. An EPRI study shows that a conventional 1 MW generator costs $267,000, so this would not cost much more than a conventional generator, and it would be far cheaper than a 1 MW wind turbine. With a conventional generator, over the life of the machine, the fuel costs more than the machine. With cold fusion, the fuel cost would negligible, so lifetime costs would be far lower.
The generator portion of 1 MW wind turbine costs about the same as a 1 MW combustion generator, but the tower costs $1.3 million. Yet wind is competitive with combustion generators because the fuel is free -- wind costs nothing. With cold fusion, the extra $119,000 you pay for palladium is far less than the cost of the wind turbine tower.
Regarding fuel costs, high purity heavy water today costs ~$1000/kg. It will be much cheaper with cold fusion, because most of that cost is for the energy used to separate heavy water from ordinary water. 1 kg of heavy water produces 69 million megajoules of heat. A 1 MW reactor consumes 3 MJ of heat per second, so that's 23 million seconds, or 266 days, or $3.76 per day. Actually, it would be far cheaper because heavy water will be cheaper, as I said.
The EPRI generator data is on p. 2-5 here: