The original heavy water plant in Vemork, Norway recovered it as a byproduct of producing hydrogen from water by electrolysis [1] for the manufacture of fertilizer. Ordinary "light" hydrogen is preferentially evolved from water when it's split by electrolysis, with deuterium accumulating in the liquid left in the cell. If the plant operator periodically extracts the residual water from a cell and uses it as input to another cell in a cascade arrangement, as done at Vemork, then high-purity heavy water can be generated as a byproduct of electrolysis with little additional cost. The Vemork plant had 108 megawatts of capacity and could produce 12 tons of heavy water per year ([1], [2]).
I expect several gigawatts of electrolytic hydrogen capacity to be built this decade for supporting "green" production of steel and ammonia with renewable electricity. Plant operators could also invest a modest effort to extract and concentrate older electrolyte batches, providing a new low cost source of heavy water. Note that this is a very energy intensive way to get deuterium, but it could be low cost because the capital and operating expenses are mostly accounted for by the plant's primary purpose (hydrogen production). The marginal cost to add a side business of heavy water recovery would be much smaller.
[1] https://en.wikipedia.org/wiki/Norwegian_heavy_water_sabotage
[2] https://en.wikipedia.org/wiki/Vemork#History