I'm assuming some plus signs are missing, and the series is

By the ratio test, the series converges for all
, since

which tells us the radius of convergence is infinite, so the interval of convergence is the entire real line.
In fact, recalling that

we see that

and this holds for all real
.