We don't have sufficient information to reliably predict IQ, but we do know hundreds of genetic loci associated with intelligence. The overall contribution of these loci is significant.
The polygenic scores predicted 4–7% of intelligence variance in independent samples; another study predicted 10.6% [50]. Thus, a blood sample at birth in these samples predicts intelligence with about the same effect size as parental socioeconomic status, i.e. they do not predict well; neither is of practical use for predicting the intelligence of an individual.
Source. (A review of the subject.)
It's true that the polygenic scores cannot reliably predict that one person will have a higher IQ than another, but that doesn't mean that polygenic screening is useless as a tool for increasing the expected intelligence of one's offspring. People who effectively screen their embryos will, on average, have slightly but significantly smarter children than people who don't. In this way, screening is not qualitatively different from many other parental interventions.
I would use this sort of screening if there was an opportunity to do so. (I don't think it currently justifies resorting to IVF if that is otherwise not necessary, although it would if the effect was larger.)