This revolutionary shift, however, is still in its initial stages, offering investors time to decide what positions they will take with regard to these changes to production, logistics, security, and numerous other industries. If you are an institutional LP who’s curious about gaining exposure to robotics ventures, you’re not alone: greater numbers of investors across both the public and private markets are taking the plunge into robotics, signaling that the time has come for long-term LP investors to take a closer look at this promising sector.
Robotics as a Service: Following the SaaS Playbook
One reason that the shift to robotics automation has been slower than some might have anticipated is that historically, the manufacture of robotics has been capital intensive. Simply put: robots are very expensive devices, and only in recent years have certain types of high-quality robots become commodities. Unlike ordinary laptop computers and servers that powered the high-margin SaaS opportunities of the past decade, robots are highly specialized machines made to suit particular tasks. A robot that unloads shipping containers, for example, must meet completely different design criteria from a robot that stocks shelves in a store, even though the two will perform relatively similar tasks. Many prospective customers at small and midsize businesses simply can’t afford or rationalize the high up-front cost of robotics adoption, even when the investment promises to pay for itself over the long term.
This resistance is giving way, however, as robotics firms incorporate lessons from successful SaaS business models. These robotics-as-a-service (RaaS) entrepreneurs recognize that the truly revolutionary potential of the robotics sector lies in the software that powers the robots—not necessarily in the robots themselves. As Paul Willard, co-founder of the robotics-focused Grep-VC and seed investor and advisor to multiple unicorn companies including Zipline, Boom, Carta, and Cobalt Robotics among others, explains, SaaS runs on hardware—and robots are just the latest hardware to become widely commoditized.
The way to speed adoption, RaaS entrepreneurs believe, is to offer a subscription-based model that allows customers to pay for robotics hardware as part of a service that also includes software-based solutions to difficult but tractable problems in their respective industries. Rather than require a hefty initial investment in hardware that may become obsolete, the RaaS model allows clients to pay for robotics services on a per-job or per-month basis. This subscription or rental fee includes supplementary support for RaaS software, as well as hardware maintenance. From the customer’s point of view, this model shortens the timeline for return on investment.
For example, Zippedi is a startup that promises to deliver “AI-driven retail.” Zippedi is introducing retailers to the cost benefits of using robotics for chores that are impractical for humans—such as creating an overnight inventory of store shelves. Zippedi’s robots roam store aisles, using computer vision to identify low- and out-of-stock items, as well as spot misplaced products. Zippedi’s robots and the software platform that operates them are able to make a store nearly 100% shelf-compliant—a task that is practically impossible for humans to accomplish—at a cost of just $100 a night. Zippedi doesn’t charge customers for the full cost of its robots. Rather, clients pay for access to a robot-enabled SaaS inventory platform. Essentially, this is a variation on the shift from the ownership economy to the rental and gig economy.
Machine Learning and COVID-19: Accelerating the Robotics Revolution
One accelerating factor in robotics adoption is machine learning. Even the most bullish roboticists admit that robots are still nowhere close to being capable of accomplishing numerous tasks that humans routinely perform with a high degree of safety and accuracy. Machine learning, however, is allowing roboticists to make rapid strides toward perfecting tasks that are difficult or impractical for humans. By compiling and aggregating massive amounts of data, sensor-equipped robots are able to find solutions to seemingly intractable problems that would otherwise require expensive and ineffective uses of human labor.
As Bowden explains, companies that approach RaaS strategically narrow the scope of the problem to a single task that is tractable within a machine learning framework—usually one involving repetitive work. For Zippedi, this task was identifying and tabulating every product in a store. The task enables store managers to increase revenue by raising shelf compliance from the human average of 90% to the RaaS average of nearly 100%. After one such profitable use case is proven, Willard explains, a company may then begin adding more features to enhance utility and bolster the value proposition. To use a different example, Cobalt Robotics, began by producing robots that patrol offices as after-hours indoor security officers. During the daytime, the same robot performs a concierge check-in service at the building entry. After proving these use cases, Cobalt began to explore other features requested by clients, such as monitoring the charge of all fire extinguishers in the building, reading carbon dioxide and carbon monoxide levels, checking for leaks, ensuring doors are locked, and other facilities-related functions. Following the outbreak of the COVID-19 pandemic, Cobalt also added a temperature check feature for visitors to the building.
Whereas machine learning has long been expected to aid robotics development, the COVID-19 pandemic injected another unexpected catalyst into widespread robotics adoption. Human labor–intensive occupations, such as those in agriculture and meat processing, have suffered some of the highest rates of COVID-19 cases. Consequently, investors and industry leaders now see the transition to robotics as both inevitable and imminent. According to Arzum Akkas, a professor of operations technology at Boston University, the pandemic has accelerated robotics adoption in the agricultural and food production sectors to address both safety and labor shortages. But agriculture is just one of the myriad industries that stand to be reshaped by robotics. One measure of global investor enthusiasm, the robotics-focused ETF index fund ROBO, outperformed the S&P 500 for almost all of 2020. At the very least, investors are convinced that the pace of automation across numerous sectors will increase as a result of the pandemic.
A Rapidly Expanding Opportunity Set
According to Market Expertz, the global RaaS market is expected to exceed $40 billion in the United States alone by the end of 2026, growing at a CAGR of more than 23% between 2016 and 2026. According to Grep VC estimates, robotics have the potential to replace 80 million US jobs, across numerous industries, including manufacturing, defense, agriculture, security and surveillance, logistics, warehouse automation, delivery, health care, hospitality, retail, construction, and entertainment and leisure.
Given the importance of robotics and the potential market size for RaaS, LPs should not ignore this sector. Throughout 2020, SaaS company stock prices and valuations soared higher as investors bet on their long-term necessity as a result of the worldwide reorganization of office work compelled by COVID-19. SaaS stocks experienced intense volatility throughout the late summer months, but overall they represented some of the biggest market winners of the world’s new business and social environments. Investors with greater insight into venture development, however, have begun to look ahead to the next generation of disruptive SaaS-supported robotics and machine-learning platforms.
To be sure, investors must place their bets carefully: successful RaaS companies are those that avoid common pitfalls that have, until recently, stymied the robotics sector. Unlike consumer-facing businesses or even SaaS software, robotics applications are complex, and the industry still lacks the necessary expertise for staffing the numerous robotics endeavors already underway. Roboticists are few in number relative to the entrepreneurs and venture capitalists with interest in cultivating the sector—most venture capitalists do not possess the requisite expertise to diligence investments, let alone offer a network of support to their portfolio companies. World-class roboticists are also highly competitive, making it difficult to assemble collaborative teams.
For all of these reasons, venture firms that specialize in robotics and prioritize assembling strong teams with deep expertise like Grep VC will be the ones to back the companies most likely to succeed. According to VentureBeat, deeptech-powered RaaS systems are positioned to “eat the world of work.” If there was ever a time to start exploring venture-backed robotics companies, that time is now.