10 Factors to Consider When Evaluating a Robotics Startup
Over the past decade, hardware costs have decreased and advances in artificial intelligence have greatly expanded the capabilities of robots. This has led to an explosion of robotics companies in recent years targeting markets such as agriculture, construction, mining, manufacturing, logistics, etc. Within industrial automation alone, 2020 robot sales hit $18B, which represents over 2.7M robots installed in factories globally.
Robots are here to stay, and are continuing to benefit from COVID-19 tailwinds as the need for functionality — without the risk of person-to-person infection — pushes more installs into both familiar and remote parts of our lives. In sectors such as agriculture, emerging technologies like field robotics are predicted to see an unprecedented amount of venture capital investment in 2021.
Since joining Toyota AI Ventures, I’ve had the opportunity to be pitched by hundreds of early-stage robotics startup founders. In the evaluation process, much of the typical criteria used by investors to appraise a tech startup at the seed or Series A stage still applies. However, I’ve found that an additional special lens is needed to truly gauge whether a robot can deliver a compelling value proposition while effectively operating in a real-world environment.
Here are 10 factors to consider when evaluating an early-stage robotics startup:
1. Is the company driven by business leaders, engineers, or a well-balanced combination of both?
Until an industry matures, the core technology often matters more than the business — initially. However, as an industry evolves and the technology becomes more commoditized, the business matters more. It’s important for the investor to recognize the market phase that the company is in before discounting what may seem like an overly weighted technical team. With that said, the team will also need to be capable of managing the end-solution where it lands. In industrial manufacturing, prior experience with systems integrators or solutions providers is critical for success in large-scale deployments. A well-balanced team has expertise in the technology, familiarity with the application domain they are targeting, and experience in being customer-oriented and delivering and supporting working solutions.
Finally, the team needs to be flexible and open to feedback. Emerging technology is new, not only for a team, but also for their customers. Visions may not always align, and it is essential to listen to the end customer.
2. Is the company’s differentiation hardware- or software-based?
There’s no question that hardware innovation is needed, but it’s often difficult for startups to find a sustainable competitive advantage. The large automation and robotics companies are able to invest in and catch up on any hardware advances much faster than software advances because that’s their bread and butter. Startups can compete better on software over the medium-to-long term since they are not locked into complex legacy systems.
3. Is it possible to achieve recurring revenue?
Most robotics startups claim to have — or are working towards — a Robots-as-a-Service (RaaS) business model. RaaS is ideal because being able to generate consistent and recurring revenue greatly accelerates the process of scaling quickly and efficiently. However, in some verticals the customers prefer a CapEx model and aren’t yet comfortable with change. It’s critical to understand how these customers think and how the product is creating value for them, and to build a strategy around converting them to a new model. Alternatively, it’s not necessarily a deal-breaker for investors if revenue does come from one-off sales, as long as growth opportunities exist and the market size is sufficiently large.
4. How well does the robotic solution work and interact in real-world conditions?
The past decade has produced plenty of examples of robots that functioned well in the lab, but struggled in the real world. Dust, dirt, temperature changes, surprising human interactions, and other unanticipated factors can cause systems to fail. For instance, food preparation machines that grip and grind squishy meats get messy and may develop mechanical issues over time due to temperature and humidity changes. Then there’s unexpected consumer behavior — such as stopping a security robot to take a selfie with it — that decreases the efficiency of the service. Inspired by cow tipping, “robot tipping” is a new issue that roboticists have to deal with when designing their machines. There are many more real-world factors that must be considered based on the environment where the robot will be deployed.
5. Is the solution simple or an over-engineered Rube Goldberg machine, which executes a simple task in an unnecessarily complex way?
Some startups design robots to solve problems without thinking through whether a robotic solution is truly merited — as evidenced by many household robots. Does one really need a $30K robotic arm to clean their dishes, or does the conventional dishwasher augmented by a human make more sense? Yes, a dishwasher could be redesigned to self-load the dishes, but the solution must be practical from both a cost and efficiency perspective.
6. Is the company obsessed with engineering an unrealistic vision driven by science fiction — or with getting customers, understanding user needs, and optimizing the product experience?
The team needs to focus on identifying problems first, not technology, followed by finding a viable business model and adapting their technology and expertise to solve the problems associated with it. It can be challenging for technical founders to avoid pursuing interesting research ideas instead of addressing what customers need or want. A team that creates and shares their roadmap early helps potential customers understand where the technology is heading and where future applications may exist in their business.
7. Is the company addressing a new white space market, or is it a fast follower to another company in a space that already has proven traction?
In many segments of the robotics industry, there’s a “first mover” company that is trying to tap into a totally new market, with “follower” companies coming later, after identifying existing business models that they will use to compete. Many fulfillment robot companies followed Kiva Systems and many full-stack autonomous vehicle companies followed Waymo.
If you’re investing in the new white space company, you have to be really confident in the founders’ vision and the expertise they bring to the table, and you have to be committed for the long term. If you’re investing in the follower company, the founders should be amazing at execution, and hopefully, they can ride the wave of the entire industry taking off. The leader often gets strategically acquired — like Kiva was by Amazon — leaving a later entrant as the new leader in the market.
8. How robust is the robot? What is the maintenance strategy?
All automated systems should have a resilient failure-recovery mechanism. Automation should augment the end-user experience and ideally allow for intervention or improvement by those who work alongside robots. Maintenance almost always requires boots on the ground to deal with issues. Designing highly modular systems and giving users replacement parts on hand can help provide continuity of service.
9. Are there any physical limitations to the technology?
Solutions should be adaptable to a variety of parameters, but physical limitations of the hardware can decrease that adaptability. Many people look at automated systems and assume that anything is possible, but when it comes to the parts, the equipment, or the systems, there are always limitations. For example, gripper robots have fundamental limitations on how heavy something can be, and a farm robot may have compute limitations on edge that limit the number of weeds killed per hour.
10. How are the robots financed?
Startups should avoid purchasing inventory with equity, unless it’s the only option available. Finding a way to set up credit lines with banks, manufacturers, and distributors can lower the cost of capital to extend a company’s runway. One creative way of lowering upfront capital requirements is by leasing out robots.
Robotics startups don’t need a perfect score across all of these factors to be investable, but the answers to this set of initial questions can help investors calibrate themselves to the strengths and weaknesses of a particular opportunity. Ideally, founders can also use these questions to understand how they might be evaluated and determine what to address in their pitch.
At Toyota AI Ventures, we have invested in several robotics startups, such as Cobalt Robotics, Elementary Robotics, Freedom Robotics, Intuition Robotics, and YPC Technologies, and we’re always looking for talented entrepreneurs who are breaking new ground. If you’re working on an early-stage startup that meets any of the criteria above, or you’re an investor looking to discuss opportunities in the space, we’d love to hear from you.
