Starship Technologies: The Future of Delivery is Robots

Source: NPR/Meg Kelly

Context: Starship Technologies is a London based startup that raised a $17.2M seed round in January 2017. Starship builds semi-autonomous delivery robots that can be used by vendors like UPS, DoorDash, etc. to automatically deliver goods to consumers. The little robots, reminiscent of WALL-E, are currently being tested in 16 countries.

I will analyze Starship’s team, market, competition, and product to deduce insights into the delivery robot space and where Starship fits in.

Team

Starship was co-founded by Janus Friis and Ahti Hendla, both of whom were on the founding team of Skype (acquired by Microsoft for $8.5B). They’re technical leaders with serial entrepreneur experience, and they’ve taken a successful product to market and exit multiple times throughout their careers.

LinkedIn Analytics also reveals cool insights into Starship’s size from a macro-perspective.

LinkedIn Analytics for Starship Technologies

We see a 3.5x growth in the amount of employees in just 2 years, with 72 employees formally listed on the network. The graph shows a fairly linear, steady growth rate, which makes sense given that Starship is operating in the hardware space. Scaling hardware takes much longer and more capital than pure software businesses.

Market Size & Strategy

The space is growing quickly. CB Insights shows more than 2x money flowing through equity deals in 2015 than in 2014, and a full-year projection of 150 total deals by end of year 2016.

Early stage deals are dominating the landscape, signifying that lots of entrepreneurs are eager to break into the space. We aren’t seeing that many companies in the later stages yet, but that’s likely because the industry is still very young.

If we break it down by sector, Enterprise takes up 48% of the total funding landscape for robot startup deals.

Starship fits into the Service, Restaurant, and Last-Mile Delivery silos of the Enterprise market. I wouldn’t be surprised to see Starship also play on the Retail & Warehouse stage, where it would compete with Amazon’s automated warehouse and other warehouse robotics solutions. Although Starship only advertises the Last-Mile use case, the tech should be portable enough to other scenarios, so I’ll treat the company as part of the other categories as well.

According to Zion Market Research, the service robotics industry was valued at $9.01B in 2016 and will grow at 17.8% annually, resulting in a $24.1B valuation by 2022. Technavio predicts a nearly 260% increase in the automated delivery market between 2015 and 2020. Starship is an early player in a rapidly expanding game, and judging by how quickly they’ve been able to get MVPs on the road (sidewalk), I suspect they’ll be able to initially capture a large part of the delivery market if they beat their competitors to market.

The strategy is especially interesting. Instead of operating direct to consumer robots, Starship is selling its bots to various businesses who have a need for delivery operations, and these businesses then use the bots to transport their goods and services. DoorDash and Postmates in the Bay Area are both experimenting with Starship Technologies bots.

The value proposition is clear: much of the overhead cost in delivering food to people comes in the form of paying drivers (usually contracted gig workers) a salary for their time. Subtracting the driver results in elimination of a variable cost in exchange for presumably a fixed cost that would pay for itself over time. According to Postmates, robotic delivery should be orders of magnitude cheaper than human delivery for the end-user.

Tailwinds:

  • “On-demand” E2E economy in full swing. Consumers want more things, sooner, and cheaper.
  • Support from VCs for automating away gig work like driving and delivering.
  • Support from D.C. in creating friendly legislation and regulation (so far).
  • Partners like Mercedes-Benz have domain expertise in autonomy and city navigation

Headwinds:

  • A new political climate (in America) fiercely opposed to tech automation and the job loss it brings.
  • Scale poses problems for city infra. If hundreds of robots are making deliveries in SF at any given moment, then sidewalks, bike lanes, and traffic will take a hard hit.

Market Competition

There’s two key competition brackets to look at here:

  1. Other land-robot delivery startups
  2. Aerial delivery (i.e. autonomous drones)

Land Robot Competition:

Because the industry is so young, there’s a considerable amount of early startups popping up in the space.

Dispatch (a16z) has completed two test runs at prominent universities of its autonomous delivery technology. Savioke is a bit further along, having raised a $15M Series A back in early 2016. Their Relay bot operates in numerous verticals, providing services in last-mile delivery, services for hotel guests and building tenants, and logistics and ops for warehouses and infrastructure. I don’t see it being difficult for Starship (and others) to pivot or even increase its product offerings to operate in various verticals other than just last-mile delivery.

Marble is in a similar boat, having raised $4M from Maven Ventures (they invested in Cruise, which had a unicorn exit to GM). Marble has strategically partnered with Yelp, so that Eat24 customers can get their food from a Marble robot.

Given that I have no access to unit economics or pricing models, it’s hard for make an educated guess on any of these companies’ trajectories. I doubt that one robo-provider will rule them all. This space seems analogous to the various car companies: they all do the same thing, some models are definitely better than others for the same prices, and yet we have hundreds of options in an incredibly competitive market. Essentially, this is not a winner-take-all game.

We’re going to see robot diversity in the upcoming years. The sidewalks will not be dotted with orderly lines of Starship robots, but rather with robots big and tall, square and round, Starship and not. Strategic partnerships likely have a lot to do with this future. Providers will partner with certain companies with some maintenance and management plans hashed out. Fleets will likely have a couple different kinds of robots, so that maintenance schedules and service tiers can be varied (i.e. super fast robots cost a bit extra than the 4mph Starship robots for deliveries).

And, just like when Costco switched from AMEX to VISA, these shifting alliances and business priorities will likely be make or break for these early stage companies, who are relying heavily on B2C’s to put the robots to good use. We’re probably in for some kind of consolidation wave where teams will band together to make bigger, stronger robot companies rather than a hundred equally weak ones.

Aerial Drone Competition:

While some of the above applies here too, aerial delivery is in a unique position to compete with Starship. It is unclear whether consumers would prefer drones to robots, or even if they care what delivers their package. I mentioned that delivery robots at scale would result in clogged cities, but drones buzzing around in municipal airspace would be annoying too.

The more likely scenario is that we’ll have robots on the ground in dense cities like San Francisco, and drones buzzing around in suburbs like San Jose. For this reason, I don’t think Starship is competing directly with drone delivery solutions. Starship makes the most sense for urban environments with chains of deliveries, that have people living closely packed together and miles of sidewalk infrastructure.

San Jose, by comparison, has bigger properties and is more spread out, with sprawling stretches of road between some neighborhoods. This makes it a bit unrealistic for robots to cover these kinds of environments, but perfect for drones which can land and take off vertically from front yards (which are all but nonexistent in SF).

Product

This little guy (or gal), has “walked” over 16,000 miles so far.

Starship Technologies has built an adorable little robot that obediently “walks” around people, dodges obstacles, and obeys crosswalks and stop signs. The product looks fantastic in video demos.

The bot is not 100% autonomous. Rather than this being a bad thing, it’s actually a feature. The robot will travel autonomously 90% of the time, but be remote controlled 10% of the time. This lets the Starship team control deliveries in cases that the self-navigation fails, bad weather, or other unpredictable situations. Even remote controlled deliveries, though, are likely much cheaper than a gig worker.

As for being a 10x product, Starship is in a prime position to disrupt human delivery workers and the service will be orders of magnitude quicker and cheaper here. I don’t know if Starship’s bots beat other providers bots, though.

This game is not zero sum, and robot diversity might be a great thing for parcel providers. In that case, Starship doesn’t exactly need to make sure it’s the best robot on the block — just that it gets the job done and for cheap.


If you want to chat about tech trends, product, or venture capital, tweet @rak_garg or email me: raksgarg@ucla.edu.