Why drone delivery is a great space for startups to be working on

TL;DR: Several hard problems to be solved for drone delivery: batteries, better GPS, and autonomous navigation in city environments. The tremendous opportunity and long list of acquirers makes it a great space for startups to work on.

I came across a recent post which argued that drone delivery was an almost solved problem and was barely months away if only FAA got out of the way. Our view based on conversations with entrepreneurs is that this months time frame is too optimistic. There are technological and systematic challenges that are pieces of the drone delivery puzzle that would take years not months to realize. The key challenges in our mind are:

  1. Battery Life: The best drones out there have a flying time of 20–25 mins, that gives them a range of 300 – 400 m. Most drones weigh around 2–4 lbs can carry say 25% of that weight in pay load (generous) which is half to 1 pound. So, today they can deliver an empty pizza box or one large meat lovers pizza from a pizzeria 0.2 miles away (too bad if you are close only to Round Table).

Opportunity: 10X+ energy density improvement of Li-ion batteries could push the delivery distance to a couple of miles when it starts to become interesting.

2. Location: The GPS in our phones has an accuracy of 20–30 m. Higher precision GPS units accurate to 5–10 cm are more expensive than average cars. So, today a drone could deliver the pizza on your doorstep or at a neighbor’s few houses away or the street.

Opportunity: Cheap hardware-based or software-based approaches to accurately determine location.

3. Autonomous Navigation: The range for remotely controlling drones is around 1 km. For drone delivery to be practical, we need autonomous drones. Autonomous flying in cluttered and unstructured city environments is a huge problem and there is no google maps yet. Drones have to not only sense objects but interpret the meaning of what they are sensing, avoid running into things, adapt to weather conditions, all of it in real time. Imagine the complexity of trade offs — the more sensing hardware you put and the more processing power on board, the less load and batteries you can carry which mean less distance you can cover.

Opportunity: Hardware for sensing, semantic mapping platforms, algorithms for real time response and several other solutions. These have to be light (weight, computational processing) and flexible enough to tackle varied city environments.

4. A robust network of drones: It took decades for the likes of UPS and Fedex to get their delivery network of trucks in place. I doubt if they would have built that network when the Ford Model T came out. Once automotive technologies matured, the puzzle was less complex. That is definitely not the case here as pieces of drone technology continues to move by leaps and bounds. I highly doubt if Amazon or UPS would build out a delivery network when the drone-related hardware and software technology are evolving by leaps.

Upside: Tremendous opportunity and endless list of acquirers of startups

Several of the opportunities above are applicable to autonomous vehicles of any kind — cars, trucks, and trains. These technologies have an outsized potential to automate ground-based logistics. The hard problems are likely going to be solved by different startups focusing on different pieces of the technology puzzle. And several large players are showing signs of entering this space. The acquirers range from hardware/chip makers like Samsung, Intel, and Qualcomm to automotive companies and their suppliers like Tesla, Mercedes, and Denso to emerging heavy weights in this field like Google, Apple, and Amazon.


We are investors in Cyphy Works.