I haven't been able to find an understandable explanation why the risk neutrality is coherent with the risk aversion implication of the expected utility hypothesis. I can see that when using the risk neutral measures it is independent of subjective choices, but I don't see the link why this is in line with risk aversions
If you make the assumption that the market is complete and that there is no arbitrage then the risk neutral measure exists which allows to price each asset as an expectation of the asset’s future value. So in utility language this means that under the risk neutral measure all agents have 0 risk aversion and their utility function is purely linear (instead of a generic concave function in other measures).
Hope this clarifies.
You can also see this other related question if you want more color