Since the 90s I am a big fan of robotics, and I become very upset when companies producing robots fail.
For the last three years a few reputable companies that I liked a lot “left the market” (Anki, etc.).
As my company licenses SLAM for robotics and AR/VR glasses developers, I understand the market from the inside and will allow myself in this article some thoughts about system problems and necessary changes in the market that may change the situation.
Why do cool robotics companies fail?
In my opinion, there are three main reasons:
1. Robots are still not smart enough to solve problems in the mass market for “mass market buyers.”
Purchases of robots by ordinary consumers are still “experimental” since there is not a single robot that solves the problem 100%
Closest to the status of “iPhone in robotics” are vacuum cleaner robots, but they are still far from ideal if judged by the video on YouTube – so many problems these robots cause in buyers’ homes are :=)
The “robots for kids” niche is undoubtedly stable – the audience of 5–8 years-old continually changes, however, in this segment, children get tired of robots quickly due to low “intelligence”, about 2 – 3 hours after the purchase, so parents buy no more than 1 or 2 toy robots, since their functionality is almost the same and they are equally dull for children).
2. Many markets with a niche audience
There are a lot of niches in the world for using robots, but there are very few places where you can find a sweet spot for a scalable business – high margins and a significant volume of sales.
Perhaps the only good example of a successful company so far is DJI, which was able to consolidate itself in the most massive segment of electronics and has been dominating the market for a long time due to its efficient sales channels and low cost.
All other startups either shut down their operations unable to sustain the competition, or focused on small segments (where there is no hope for business growth), or pivoted into the development of hardware or software components.
3. Large investments in the development of basic components of robots due to the lack of ready-made components
To create a new product for any robotics startup, you need to create a considerable number of components and software from SCRATCH – there are simply no ready components of software of the expected quality.
Another headache is their integration needed to achieve the desired functionality.
According to my statistics, more than 50% of startups in robots get stuck during the creation phase of the product, facing the need to create missing components (the creation of which is often itself a separate fundamental task “worthy” of another startup).
The combination of factors 1-3 ultimately leads to a high cost of the final product and low demand from buyers who do not receive proper solutions to their problems at the appropriate level.
When will awesome robotics companies stop failing?
At a certain level of abstraction, any consumer robot a physical product is a part of a mobile phone (same CPU, Memory, IMU, OS, etc.) with additional physical components (wheels/propellers/ engines) and unique AI robot control systems.
Since the phone elements are already extremely cheap (due to savings on the scale during the production of components for phones), to achieve the minimum cost of robots and a short development cycle, the market lacks two more elements – “on the shelf” AI systems (SLAM, , etc.) and hardware (wheels, engines, propellers, etc.), which could be integrated and operated as cheaply as possible.
How will this problem be solved? I expect that the market consisting of fully integrated robotics companies that are trying to do everything in-house awaits “LEGOfication”; instead of vertically integrated companies, there will be many different highly specialized companies that will develop and deliver various puzzle elements (from “AI brain” to “propeller” and” wheels”). This will allow a large number of OEMs specializing in different robots industries to be created (to create specialized robots, they will need to JUST assemble them from the components instead of going through the expensive development process).
Understanding the prospects of such a model, manufacturers of chips and autopilots have started entering the market.
The second driver of the market are startups, which instead of producing hardware decided to pivot and make AI solutions for robots (SLAM, maps, path planning and so on).
With current dynamics, I’m pretty sure that in 3-4 years the quality of all these puzzles will reach a level that will allow us to very quickly create intelligent robots that will effectively solve specific problems of end-users!
Best regards,
vactivity @ gmail.com