My first reaction was somewhere between “huh?” and “you’ve got to be kidding.” After all, every box would need to be either waterproof or put into a watertight container. And each container would be fitted with a device that could communicate with a network for retrieval. More problematically, the device/container has to be capable of sinking to the bottom of the pool for storage and then rising to the surface on command to be pulled out and put on a delivery vehicle. Extra steps, extra materials, and more cost all around.

But e-commerce is exploding, and with it, the need for storage and space to stage goods and quickly get them to customers. So there are some interesting aspects to this proposal. For one thing, each individual item moves itself as the container/balloon and box expands or contracts and the product floats up or down in the water, so all or most of the handling goes away in the ideal setting. The whole process of moving goods around the fulfillment center takes less energy because the water supports the box, and currents in the water push the goods, just like barges take less energy than trucks. Finally, think of the flexibility you get. If you need more space, you don’t build walls; you fill up what amounts to a wading pool and just back up conveyors to both or all sides. And since many big cities are close to water, you might actually be able to store the goods near your customers.

Don’t get me wrong. I think actually doing this would be cost prohibitive because of the containers, which have to be handled both for storing and for unboxing, as well as either the labor or the mechanization involved. And I’m not sure what you do when the box you need is the one on the bottom of several layers, rather than on the top. But the ideas involved do hint at solutions to the “last mile delivery” dilemma.

First, having every item move itself (at least up and down in the pool) eliminates much of the handling that currently goes on in fulfillment centers, especially the search effort. This is the ultimate “goods to people” configuration, and it’s driven by the item itself.

Similarly, storing goods in a medium which moves, i.e. water, eliminates aisles and all the other empty space that humans, and even robots, require. So the dream of complete space utilization gets closer. We already have movable shelves, kiva robots, and other automation, but if you could somehow use nature to move things horizontally as well as vertically, you might be miles ahead.

In fact, this whole idea is another example of Amazon’s continuing interest in more efficient logistics and warehouse automation, no matter how far “out there.” Amazon’s current fulfillment centers are stuffed with robots that move goods from dense storage stacks to people who make up the orders that you and I receive. Predictive analytics is at play as Amazon tries to buy the minimum needed to satisfy customers. And there are reports that some Amazon vans are equipped with 3-D printers so that common items are literally manufactured en route to the customer.

All of us in the logistics business admire the relentless drive to improve what seems to be part of the Amazon DNA. On the other hand, many of us wonder about the balance between profit and investment that goes with ongoing capital investment and innovation.

In any case, Amazon (AMZN) and others continue to look for new ways to make e-commerce “work,” i.e. very fast delivery with close-to-free shipping. And there is an “end state” I suggest to my students we can all aspire to: Tell the computer what you want, and it materializes.

Oh, that’s right, somebody already thought of that. But until we can dial-up Star Trek, any initiative that addresses labor productivity, space productivity, and energy efficiency of movement, no matter how out there, is welcome food for thought.

Arnold Maltz is associate professor of supply chain management at Arizona State University.