Within the robotics sector, a few big corporations are accelerating robotics adoption by investing in and acquiring small teams of developers and using their economies of scale to drive the industry forward. Since 2012, Amazon has gone on a spree, acquiring mobile robot developers to power the automation of material handling. On a smaller scale, Teradyne has decisively picked winners in emerging markets to cement its future as traditional revenue streams in electronics testing stagnate. Likewise, FLIR is in the process of integrating a number of military robot companies into a significant roster. But perhaps Softbank, the gigantic holding company and possession of Masayoshi Son, is the most important private actor in the robotics space internationally (see Table 1, below).
Bear Robotics and Berkshire Grey
2020 has seen two significant investments so far. One US$ 32 million investment in mobile hospitality robot Bear Robotics, which is being targeted for restaurants and casinos, and a significantly larger US$ 263 million investment in Berkshire Grey which develops automation solutions for retail, e-commerce, and logistics fulfillment operations. According to Berkshire Grey, the new funding will help fund its global expansion, acquisitions, and team growth. The company’s robotics solutions aim to accelerate the transformation of customers’ logistics operations to address consumer expectations, labor shortages, and competitive pressures.
Zume and Cloudminds
These two investments showcase Softbank’s continual excitement for the robotics opportunity, even in light of difficulties with previous recipients. For example, Zume, which received US$ 375 million in funding in 2018, recently cut half its staff and switched completely from its initial focus on automating Pizza production to developing third-party pizza delivery and packaging, with the robots. The company is noted to have lost US$ 50 million and has been characterized by lack of focus, although it retains the majority of Softbank’s funding. Furthermore, Pepper, the famous social robot developed by Aldebaran and acquired by Softbank in 2012 may have sold in large quantities, but has done little to push industry adoption forward, with many end users not renewing their leases of what is ultimately a gimmick.
Other cases are more nuanced. Artificial Intelligence solution developer CloudMinds have raised a reported $317 million in total VC funding from investors, including SoftBank’s Vision Fund, which holds a 34.6% share in the business. SoftBank previously participated in a seed round for CloudMinds and reportedly backed a $186 million funding for the company earlier in 2019. The company has recently been in the news for deploying 5G-enabled robots for disinfection and medical delivery use-cases in Wuhan and Shanghai in the aftermath of the Coronavirus outbreak.Softbank’s
CloudMinds is affecting the robotics market with its enablement of data processing, advanced vision and smart voice competencies for the future robot fleets to be deployed. The company logged a 2018 revenue of $121 million but with a net loss of $156.8 million, compared to a 2017 revenue of $19.2 million on a $47.7 million net loss. This is further evidence that short term profitability is not high on the agenda and that Softbank is willing to take the long view on robotics.
The value of Pepper will deteriorate further, as the social robot opportunity is decidedly overrated in its value and will not have a significant impact in the near term.
Uber’s Gordian Knot
Unfortunately, similar bets with We Work have likely failed. This has not stopped Softbank making an enormous US$ 1 billion investment in serial loss-maker Uber’s advanced technology group (ATG). The ride-hailing company lost a staggering US$ 8.5 billion in 2019 and had to lay off 1,000 employees earlier in the year. Whilst the company managed to increase revenue from US$ 11.3 billion 2018 to US$ 14.1 billion in 2019, investors are increasingly becoming frustrated with its inability to chart a course to growth.
Autonomy might be the way out of this Gordian knot. In February 2020, Uber was granted a permit to test autonomous cars in California after suspending activity following the fatal killing of a pedestrian in Arizona, 2018. The ride-hailing company had remained testing in Pittsburgh and sees autonomous ride-hailing as the route to long-term profitability (the company hopes to turn a profit by 2021).
Robo-taxis may help them get into the black, in that they get rid of Uber’s much contested compensation to its workers, but the idea of large-scale ride hailing is unlikely to materialize by 2021. The hope to shift ride-hailing into an autonomous age is coming at a time that Uber is being constricted by ever-tighter regulations, including workers rights laws in California and losing their license in London. But the road to automation of ride-hailing is itself strewn with regulatory barriers, and the idea this can be scaled up by 2021 is fanciful.
If autonomous robot taxis are Uber’s master plan to profitability, it will be a very long decade of maintaining the faith of increasingly impatient VC’s while fighting against the company’s deteriorating image in the eyes of activists, municipal authorities and political parties across the ideological spectrum.
Some Success, But Pressure On
The pressure on Softbank’s beneficiaries are also felt by the company itself, which saw a US$ 2 billion reduction in profit in 2019. Many of Masayoshi Son’s own investors, including Paul Singer’s Elliott Management, are increasingly demanding greater discipline and stringency when it comes to making sure returns on investment are shorter than they currently are.
This is not to say Softbank have not also picked success stories. All the companies mentioned have impressive technology and most are reporting strong revenue growth. A particular outlier is Brain Corporation, whose installed base of 6,000 autonomous sweepers stands out as the biggest deployment of autonomous commercial vehicles outside of manufacturing and logistics.
Brain Corporation has already deployed at impressive scale and will likely expand on this with the rollout of Softbank’s Whizz bot (which uses Brain OS), while Berkshire Grey and Bear Robotics have real potential to scale thousands of shipments without facing significant regulatory barriers or exorbitant upfront costs.
Scattershot Investment, But Some Will Pay Off
So, are the robot-related investments more akin to Softbank’s successful investment in Alibaba, or more like the investment in We Work, with skyrocketing losses and misplaced hype leading to an abrupt crash? There are likely to be many casualties. The value of Pepper will deteriorate further, as the social robot opportunity is decidedly overrated in its value and will not have a significant impact in the near term.
The Zume investment is not being returned with deployment of end-to-end automation of pizzas. The prospect of using articulated robots for fresh food production was risky and has led to a significant pivot away from the initial high-tech focus towards a more mundane proposition, with accompanying losses and a swift reduction in staff.
The ownership of Boston Dynamics could be rewarded with increased use for industrial inspection, but 2020 will tell if deployments for quadrupeds can be scaled beyond testing. Uncertainty is a common entity in the world of venture capital.
When Softbank brought Boston Dynamics off Google in 2017, the robot developer was exceptional for developing bipedal and quadrupedal robots (primarily for marketing purposes). Now, as they enter commercialization, there are some impressive new challengers in this space on similar trajectories and hoping to offer a low sales price, including Agility Robotics, Zoa Robotics, Ghost Robotics and ANYbotics.
Autonomous Driving Investments
In terms of pure sunk cost, the key application for Softbank is autonomous driving. Exceeding US$ 3 billion, the investments in Nuro, Uber ATG and Cruise may well offer strong returns in the long-term, but will likely require additional support and a delay of at least 5 years.
Outdoor road-based autonomy suffers from huge technical and regulatory challenges. Nuro’s investment is of particular note because of the aforementioned exemption and because it offers a transformative platform. If Nuro can assure regulators that driverless vehicles with no manual operability safely works, the potential dividend is enormous. But as stated, this is a long-term opportunity.
Large automakers and tech companies have acquired and invested in self-driving start-ups to the tune of billions of dollars. If self-driving passenger vehicles are taken out of robotic investment calculations, the investment figures become much less impressive. But whilst these big corporations understand the potential, they are by and large expecting this to be a very long and drawn out transition, a far cry from Google’s projections of driverless cars by 2009. Automotive companies are in fact likely to prefer incremental assistive automation over full autonomy because they do not wish to disrupt their position in the industry.
The Most Promising Investments
The most promising investments are not the most experimental technologies (Boston Dynamics, Pepper or Zume Pizza), nor the massive investments in self-driving road-based vehicles (Uber, Cruise, Nuro), but rather the mid-range investments in mobile robots for industrial and commercial applications, notably Brain Corp, Berkshire Grey and Bear Robotics.
Brain Corporation has already deployed at impressive scale and will likely expand on this with the rollout of Softbank’s Whizz bot (which uses Brain OS), while Berkshire Grey and Bear Robotics have real potential to scale thousands of shipments without facing significant regulatory barriers or exorbitant upfront costs. Whilst CloudMinds is operating at a loss now, it is well-placed to serve the expansion of AMR fleets over the next 2-4 years and is already scaling up quickly.
When determining where Softbank, and other VC’s, are likely to find success in robotics – it is mid-range investments in companies who can deploy mobile robots today.
As part of the Strategic Technologies research team at ABI Research, Rian Whitton provides consulting and analysis covering robotics, automation, intelligent systems, artificial Intelligence and machine learning. He has also written actively on the commercial application of unmanned aerial vehicles. Whitton has also been an editor and contributor to the Conflict Comment, a media outlet that focuses on international relations. He holds a Master’s degree in Science & Security from King’s College London, where he focused on the intersection of technology and defense. Subjects of interest included lethal autonomous weapons, unmanned systems, aerospace innovation, nuclear deterrence and how it relates to international affairs.
- Tethys Enables Students to Code for SoftBank’s Pepper Humanoid Robot
- SoftBank Whiz Cleaning Robot Available Outside of Japan
- Waymo Raises $2.25B in First External Funding Round
- Simbe Raises $26M for Tally Retail Robot, Expands SoftBank Partnership
- Nuro Raises $940M from SoftBank to Expand Self-Driving Delivery Vehicle Reach
- How SoftBank Robotics Builds Human Trust When Building Pepper