If I recall my geotechnical history, piles were loaded to 2X design load, not simply to provide a safety factor of 2 but to statistically assure that all production piles have a very high degree of certainty of having a capacity at or exceeding the design load.
If a pile load test is considered the "sample" and the double design load represents the mean of the capacity of all piles, then (assuming a normal distribution of pile capacities) the risk of any one pile on the project having a capacity less than the design capacity is statistically very small.