Pork. It is a jobs program not a technology program.
The SRB's are built by ATK, but the design ensured they needed to be reworked, so they could not just the existing Shuttle ones. Need an extra segment, which in SRB land means redoing the inner contour to control the burn rate to control thrust profile.
The switch between SSME to RS-68 to SSME (Yes, some of that was Ares-V, and some was SLS but basically same ideas) wasted a lot of time and money, but also meant that the SSME's could not simply be used. They needed a new digital controller upgrade. Which is probably good, but costs more in development costs.
The main barrel section is designed based on the ET from the Shuttle, but of course, needed new tools to build it since it is apparently just enough different.
The upper stage engine has been a total goat rodeo of indecision. J-2X, RL-10's, SSME, can someone please decide and stop switching. Each time money is spent, time is wasted, and little comes out of it.
In reality the true cost is in salary. So many people are employed in working on this, in by design inefficient fashion, so as to keep as many jobs as possible every time delay costs millions if not billions in salaries.
The reason other companies can launch for cheaper is that they employ fewer people. Raw material costs for launcher and fuel costs are a pittance. They are not manufactured out of solid gold. (By weight, simply building an SLS out of literal solid gold, might be cheaper than the current project. Anyone want to do the math?).