ABSTRACT Ultramassive white dwarfs with masses $M\gtrsim 1.1\, {\rm M}_{\odot }$ probe extreme physics near the Chandrasekhar limit. Despite the rapid increase in observations, it is still unclear how many harbour carbon–oxygen (CO) versus oxygen–neon (ONe) cores. The origin of these white dwarfs and their strong magnetic fields – single stellar evolution or a stellar merger – is another open question. The steep mass–radius relation of the relativistic ultramassive white dwarfs shortens their crystallization time $t_{\rm cryst}$, such that the recently proposed crystallization dynamo mechanism may present an alternative to mergers in explaining the early appearance of magnetism in the observed population. However, the magnetic diffusion time from the convective dynamo to the white dwarf’s surface delays the magnetic field’s breakout time $t_{\rm break}\gt t_{\rm cryst}$. We compute $t_{\rm break}(M)$ for CO and ONe ultramassive white dwarfs and compare it to the local 40 pc volume-limited sample. We find that the breakout time from CO cores is too long to account for the observations. ONe crystallization dynamos remain a viable option, but their surrounding non-convective envelopes comprise only a few per cent of the total mass, such that $t_{\rm break}$ is highly sensitive to the details of stellar evolution.