Functionality X at $500 has a very different market than the same functionality X at $1500; if a startup is aiming to check for product-market fit, then the expensive manual domestic builds would be aimed at a different target market that also requires and uses different features, requires different marketing and sales channels, etc with requirements and preferences that often are entirely opposite of what the "real" product should have.
Such an iteration would help in handling the actual manufacturing, but would not help in building all the other parts of the business as intended as it would drive all of them in a wrong direction.
Furthermore, who funds the very large fixed costs of design and development? If they had chosen "more manual domestic builds in the low hundreds of units" then they would have saved money/scale in manufacturing but would not have the funding needed for the engineering team to make the actual product, whatever they had was funded essentially by pre-sales.
It's also reasonable to have products that are totally unviable if the volumes are low, because to get an adequate price you need economies of scale. A business plan that clearly states "we either go big or go bust" can be a valid business plan; many real industries have a situation where selling just 1000 units is worse than selling 0 units, you either manufacture at scale or close up the shop.