I read that 10% of autistic people and 1% of nonautistic people are savants. It seems like lacking an ability is what defines somebody to be autistic.
Disclaimer, this is less of an answer and more of a critique of the demographics behind Savant Syndrome.
Currently there are very few issues with finding a decent answer to this question: This 10% of autistics being savants seems tenuous, there is no clear cut unambiguous research from current and respected journals analysing the prevalence of savant syndrome. The last paper to back these claims was in 1988, and that is the only one reporting these numbers. This 1988 paper seems to be to source of the statistics reported in the question and on websites such as the autism institute. However there have been no repetitions of this reported prevalence. Savant syndrome is a popular idea that people with learning disability can be brilliant too, and they can, there are real cases of Savant Syndrome. A few journal articles describe the workings of Savant Syndrome, such as the brain function and abilities. However these are not common, much like Savant Syndrome, if Savant Syndrome was as common as reported by the 1988 paper we would see savants amongst autistics as we do left handers across Europeans. From personal experience I would say we do not, I spent many years working with people who have ASD, I meet any savants, even though I would have liked too. Of course it is entirely possible that I could I just happened to be statistically unlucky. The lack of further prevalence data is suggests that these results haven't been replicated, and may therefore be misleading.
So whats causing this statistic? In my opinion its likely to be two factors. First the over diagnosis of autism, many children are diagnosed with Asperger’s who are not later in life. Secondly intelligent children are often diagnosed as autistic or with mood disorders, intelligent people may think differently to the general population and some of their tendencies may seem 'autistic', like there fascination with particular subjects/objects, such behaviours might contribute to a ASD diagnosis in children.
There is no prevalence data for savants in 'normal' populations, but you could say anyone who is talented and passionate about a particular skill or subject is a savant, from footballers to scientists, many 'normal' people are exceptional. Personally, I work in an academic departments, and I meet people that would fit the definition for the abilities of a savant on a daily basis, but they don't have a disorder. You probably do as well, yet because they don't have a disability you wouldn't use the same term savant, and technically you shouldn't.
Conclusion: More data needed! Their clearly isn't enough support to make the claim that 10% of autistics are savants, and additionally there seems to be no real comparison with 'normal' functioning individuals. And the definition of savant is also dubious, can normal people be savants?