It is an amortized cost. The NYT claims Medicare costs the government roughly $6,000/person/year. So the question is whether the motorcycles reduce an average of 3 years of life per person. (Ignoring price shocks from giving them out.)
But... 2015 data says ~5,000 cyclists died out of a total population of 8,600,000, making for .05% of cyclists dying per year. Assuming a grant at 65 and a life expectancy of 85, that's an average of .01 years of life lost in the first year. (And this assumes everyone rides them as much as people who choose to buy motorcycles!) The years lost decline as people age, of course, but even raw multiplication only gets us an average of .2 years of life lost per person over the course of the project.
Obviously this doesn't adjust for varying risk profiles compared to normal motorcycle riders, but we'd be talking about being more than 15 times as likely to die as most motorcyclists to make this math add up.
All moral issues aside, it looks like the math just doesn't back this one.
https://www.nytimes.com/2014/09/23/upshot/medicaid-gives-the...
http://www.iii.org/issue-update/motorcycle-crashes
edit: taking bike price down to $8k doesn't really change this logic, especially given the "not actually riding" issue I didn't factor in.