Sand casting still exists for good reasons
An ancient process that remains the right answer for certain parts.
Sand casting still exists because a reusable metal mould only pays for itself once a shape is going to be made hundreds or thousands of times, and for a large, one-off or low-volume part, packing loose sand into a temporary mould remains faster, cheaper and more flexible than building a permanent tool that might only ever be used once or twice.
Buckets for turrets, hands for walls
Building a sandcastle at the beach makes the trade obvious to anyone who has done it. A plastic bucket gives a perfectly repeatable turret every time it is packed and turned out, the advantage a permanent metal tool offers for a part made thousands of times. For the walls, the moat and any larger or odder shape around those turrets, packing and carving wet sand by hand is far quicker, because making a dedicated bucket for a shape wanted only once would cost more effort than it could ever save. A foundry choosing between sand and a metal die is weighing the same thing, only with far more money on the bucket side of the scale.
A mould that is broken open every time
A permanent metal mould, the kind die casting depends on, is expensive to design and machine because it has to survive thousands of cycles of molten metal injected under pressure without wearing or losing its shape, and that durability is what the buyer is paying for. Sand casting packs sand tightly around a pattern of the desired shape, lifts the pattern out to leave a cavity, pours molten metal in under nothing more than its own weight, and breaks the sand away afterwards to reveal the casting.
The mould is destroyed every time, which sounds wasteful next to a tool that lasts for years. It also means there is no precision-machined mould to design, pay for and wait on before the first part can be poured. A pattern, usually far simpler and cheaper than a metal tool, is the only real investment standing between a design and a finished part, and the sand itself is largely reclaimed and packed again for the next mould.
Big parts, small runs and changing designs
Sand casting is the natural fit wherever a permanent tool's upfront cost cannot be spread across enough parts to justify it. That covers large, heavy castings such as a machine tool's base, an engine block prototype or a bespoke one-off part, where the sheer size of the piece would make a metal die enormously expensive even before the low volume counts against it.
It also suits a design that is still changing. A new sand mould can be built around a modified pattern in time for the next pour, while a metal tool locked into a fixed cavity has to be reworked or remade every time the design moves, a cost nobody wants to pay while a part is still being developed and proven.
So the useful question at the concept stage is which process's fixed and variable costs match the number of parts the project needs. A permanent tool has a steep initial cost that spreads thinner across every additional part, and it only drops below sand casting's flatter, pay-as-you-go cost once enough parts have been made to absorb that investment. A project that stops well short of that crossover is paying for durability it never uses, and sand casting lets it avoid that trap at some expense to surface finish and dimensional precision.
When volume climbs past the crossover
Once volume climbs high enough, sand casting's own per-part costs, the sand handling, the pattern wear and the slow cycle of packing and breaking a new mould for every casting, keep accumulating in a way a permanent tool's running cost does not. A product that starts in sand and grows into high volume usually benefits from moving to a permanent tool at some point, even though the switch means finally absorbing the tooling cost that sand casting had, until then, let it avoid.