Feasable, and the concept has been proposed, but doesn't look likely to be built in the near future. There are still lower hanging (more profitable) fruit when it comes to building undersea HVDC cables.
It's not carbon free. Iceland's geothermal fields have carbon emissions because gasses trapped beneath the surface are released along with the steam when they're extracted. It's still low-carbon compared to a natural gas power plant, of course, but not compared to wind/hydro/nuclear.
And aluminium production is certainly not carbon free: the smelting process reduces aluminium oxide to aluminium metal using carbon electrodes, producing around 14 tonnes of CO2 per tonne of aluminium.
The point is that smelting the aluminium takes tons of electricity, so doing it in Iceland where that's produced via geothermal is effectively exporting that electricity.
It's actually the reducing of the alumina (aluminum oxide) to metallic aluminum that takes huge amounts of electricity. And as mentioned, that is done with carbon electrodes which are consumed in the process, leading to relatively high CO2 emissions. Though yes, if that electricity would be produced by burning fossil fuels the emissions would be even higher. So it's not like there aren't big benefits to doing aluminum refining in Iceland, or other places with low-emission electricity.
There is some R&D work going on though to do this reduction step without CO2 emissions using other electrode materials, see e.g ELYSIS.
And it’s a relatively light material. So if you’ve got some place where the carbon footprint of collecting and transporting bauxite is relatively low, you can use excess power to smelt more aluminum.
The problem with opportunistic loads like wind and solar is whether you can afford to strand expensive factories full of equipment for hours or days at a time while the power availability is compromised. At least with geo this is a smaller problem.
We do have the technology to build HVDC cables from Iceland to Britain / Norway and we can expect the loss of this grid-to-grid interconnect to be < 5%. It's a different question entirely if it is feasible. It would be the longest sub-sea power cable ever, and the projected cost of $4 billion might be much too low.
In the current situation Europe would profit immensely by sending excess renewable energy to Iceland's pumped hydro and Aluminium smelters while using their geothermal baseload capacity. But in 15 years that might no longer be the case and by then the investment would not have paid off and there might be regret that the money wasn't spent on a different HVDC line like another North Africa - Europe link or Bulgaria - Caucasus (which has a lot of undeveloped hydro potential).
https://en.wikipedia.org/wiki/Icelink