I think the Minnesota Transracial Adoption Study [1] is the most interesting study on this question. It was intended to be a study that would, once and forever, put to rest any question of race and intelligence. You had numerous well-to-do white families, with mean IQ a sigma above the mean, adopt children from a variety of different races. The study then tracked these families and their outcomes while working to ensure relative balance in education, opportunity, identical testing standards, and every other variable they could reasonably control.

However, in the end there was a 18 point IQ difference, at age 17, between the adopted white children (105.5) and the adopted black children (83.7). Half white/black children fell almost exactly in between (93.2). The study also had some interesting accidental (?) control variables in that some children had been racially misclassified, but their IQs ended up aligning with their race rather than their identity.

Of course one can still argue that this is environmental, by appealing to e.g. prenatal or social biases and the like, but I think there is no evidence based argument that there is no difference between races, even when every effort is made to eliminate as many viable environmental factors as possible. Obviously the mean doesn't define the individual. There are plenty of high IQ black individuals, and plenty of low IQ white individuals. But group differences are nonetheless very real.

[1] - https://en.wikipedia.org/wiki/Minnesota_Transracial_Adoption...