[potatolicious]

Combining my replies:

> "Sounds like an opportunity for a standardized, flexible industrial robot platform geared towards rentals."

This would work if different businesses have significantly different seasonalities - but on the aggregate in North America this is not true. Amazon gets the same Xmas rush as Wal-Mart, along with Target, Macy's, and whatnot. The number of businesses whose rush season is out of sync with Xmas is quite low. In fact, on the whole, retail basically rises and falls all at once throughout the year.

> "Why not go one step further and have elastic fulfillment centers?"

AFAIK Amazon already does this :) Look up "Fulfillment by Amazon".

[rabidsnail]

> AFAIK Amazon already does this :) Look up "Fulfillment by Amazon".

So they do! It looks really expensive, though. Like more than $2 for a t-shirt. I wonder how that compares to doing it yourself at scale.

[onemoreact]

I don't think China, India, and the US have the same 'black Friday' so if you ship them at reasonable cost you can probably do some load balancing internationally.

[marshallp]

But with robots, you can probably get them to be cheaper even when taking into account seasonality. The real problem is that lots of manufacturers/wareshouses only keep a 3 year investment horizon. If they invested with a 10 year payoff, a lot more automation is possible. (It mirrors the problem in the economy as a whole - short termism)

[potatolicious]

> "But with robots, you can probably get them to be cheaper even when taking into account seasonality."

I can't say much without violating some NDA or another, but I'd check that assumption. Industrial robotics are anything but cheap, and remember, you're stacking them up against near-minimum wage laborers who have little to no benefits and you only pay for them when they're utilized.

[dkersten]

Why not combine the two? Robots all year round and then hire short term laborers for the seasonal peak times.

[potatolicious]

Quite possible, and the hybrid approach is practiced quite widely - e.g., robot packing machines, robot sorting machines, etc. One big issue with hybrid approaches that require robots to run around warehouses is that humans are squishy and present a significant safety hazard that will also reduce efficiency. Much of a totally automated solution's appeal is the dramatic efficiency increase that comes from not having squishy humans on the playing field.

[vellum]

You could make separate areas. One for the robots, one for the humans.

[moheeb]

You mean have Robocop and Darth Vader stacking boxes?