How can an A/B test show significant result without enough data

I am running an A/B test and to check if I have enough data I use this calculator which tells me that it does not have the required sample size per variant, but does tell me that variant 1 is significantly worse with p = 0.0387.

I assume that you can’t tell if A or B is significantly better or worse if you did not reach the required sample size per variant. Isn’t these two metrics linked to each other?