I obviously don't know the full answer, but I strongly suspect it is combination of any of the following engineering and practical issues:
- Reliability in less than ideal weather. Drones are light so easily tossed around in the wind. And does it work at night, or in the rain when the cameras don't see so well? In many poor-weather markets that means having to say 'sorry' quite often.
- Obstacle avoidance. People, cars, flag poles, Power lines and trees near customers homes need to be reliably mapped or dynamically spotted using on-board sensors. This is challenging: fully autonomous requires more sensors.
- Basic reliability: how many fail to find a safe landing site? How many drones crash due to mechanical failure. How many will still fly into trees or power lines? The success rate needs to be way above 99%, and that is not what current drones have.
- Range and endurance. Physics is a bitch: Energy demand is high and the drone needs to fly round-trip, resulting in a range of ~10 miles tops. Larger battery means more weight, so its a catch-22.
- A huge investment in a fine-grain drone airbases and warehouse infrastructure to serve relatively few people in suburban settings. This also means that only a few popular products are in stock for drone delivery.
- Limited payload weight and size reduce practicality. Drones can deliver an Echo Dot, but not a regular amazon Echo. Pizza boxes are too large. If the drones get heavy, the possibility of people getting killed goes up significantly.
- Limited products. How many high-volume, small, low-weight, sub-$100 products exists that you need to be delivered in less than 1 hour? Likely answer: Too few to make it feasible.
- Cost of drones and infrastructure, inevitable losses, worn-out batteries, maintenance crew. This likely makes the delivery costs of a $40 Echo Dot higher than the value of the product.
- Other practical stuff: Is this only for suburbanites that have a garden? What do customers do to prep the landing site? How would you sign for delivery higher value products?
- Limited gain vs a conventional delivery van, even if all of this works. Amazon already does same-day delivery for many goods already. So all this trouble to go from 6 hours to 1 hour? And if there is really a need for such very fast delivery, why not employ delivery boys just like pizza places do? They have a bigger range, can deliver anywhere anytime and can deliver much heavier packets. Domino's can do it in 30 minutes!
Some of the above might be solved, but any of the above is a potential show-stopper. And it makes even most niche application really unlikely
I dont particularly disagree with any of the above. Obviously the technology isnt already perfectly solved or we wouldnt be having this discussion because Amazon would already have their drone fleet.
On the other hand, I dont agree that "any of the above" are show stoppers. Youve got a lot of points in there that are about practicality and market, not about technical feasibility. The fact that it would be a limited range of products or benefits over a conventional delivery etc arent indications that Amazon cant deliver, they are just caveats that Amazon may have to loss-lead or to rely on the cool factor of drones rather than immediately having a useful and profitable product. To me, a drone delivery service that runs at a loss and is offset by the advertising value doesnt count as some form of "quackery" so long as the drone delivery is an available service that does what it says.
The weather reliability and small obstacle avoidance are genuine issues. Thats an engineering and software problem. Im just not confident that some combination of development and re-defining the use case cant overcome the issues. There are issues, I just wouldnt go anywhere near saying there are issues so great that claiming to be able to fix them is quackery.