Solid helium is unusually compressible for a solid, so the underwater pressure would compress it into a smaller sphere. That heats it up, presumably turning it back into a gas.
The fact that it is very cold does not mean it freezes a lot of water. Like in the question about the 0 K iron cube in the living room, the thermal capacity of water is much bigger than the thermal capacity of the solid helium and there is far more of it. The pressure is definitely high enough to make even a thick spherical ice shell buckle and implode immediately.
So the sphere implodes and heats up, and the implosion becomes turbulent, mixing the water and the helium which increases the heat transfer. At this point things become very complicated as I would expect some explosive bounces, but in the end it turns into supercritical liquid (since at the 108 MPa pressure in the Marianas Trench at a few Kelvin this is the equilibrium state; see the phase diagram). Note that this is a normal (critical)liquid rather than superfluid.
The bubbles of helium start ascending. They quickly reach terminal velocity, which is $$v_t = \sqrt{\frac{4g(\rho_L-\rho_b)d^2}{3C_d \rho_L}}$$
If we use $\rho_b\approx 125$ kg/m$^3$, $\rho_L=1000$ kg/m$^3$, $d=1$ cm, $C_d\approx 1$ we get $v_t \approx 0.03$ m/s. At that rate it takes about 102 hours to get to the surface. However, at the start the helium will be compressed and later expand, there will be big turbulent blobs and whatnot producing currents, likely making a plume dragging water with it, so the actual speed is likely a bit higher.