It’s admittedly been a while since I’ve looked at it, but I recall that the creation of H2 using existing processes takes quite a bit more energy than you ultimately extract back from it. For example, electrolysis of H2O to H2 is a pretty lossy process (I think around 60-70% of the input power is used to break the H2O bond and the other 30-40% is released as heat).