Robotics scene continues to be bullish, but layoffs are looming

This startup season is filled with goals of profitability, promises of higher margins and whispers about pivoting toward sustainability. So when it comes to robotics, a capital-intensive sector that has a longer sales time horizon and loads of infrastructure hurdles, tensions feel inevitable.

Or at least, you’d think. Crunchbase data shows that, despite a creaky market, venture funding for robotics startups remains strong. It’s a dissonance worth exploring, so that’s exactly what we did at TC Sessions: Robotics 2022 with investors Kelly Chen, partner at DCVC, Bruce Leak, founder of Playground Global and Helen H. Liang, founder of FoundersX Ventures. The trio of investors spoke about how the ambitious sector is surpassing some of the downturn’s harshest symptoms.

The answer includes a shift in investment strategy and Amazon.

No more moody robotic arms, please

Tortoise co-founder Dmitry Shevelenko: ‘You can’t do too many things at the same time’

For a company named after a slow reptile, Silicon Valley startup Tortoise has made some quick pivots into new business models over the past year.

Co-founded in 2019 by ex-Uber executive Dmitry Shevelenko, the company began with a mission of being the operating system for micromobility vehicles, one that uses remote operators to reposition shared electric scooters to locations where prospective riders are or send them back to the warehouse for a charge.

In January 2021, Tortoise began working with shared micromobility operator Spin to test three-wheeled scooters embedded with Tortoise’s repositioning software.

But right before the company scored its Spin pilot, it started realizing the potential behind remote positioning and all the cameras and sensors the company had placed on scooters. With COVID-19 causing the burgeoning shared micromobility industry to take a nose dive at the same time as people, huddled indoors, began to demand quick delivery services, Shevelenko realized it “would be malpractice” not to pursue the robotic sidewalk delivery.

Tortoise started delivering with smaller local clients first, and then with big names like grocery story chain Albertson’s, nationwide logistics company AxelHire, and convenience store chain KRS. All signs were pointing to sidewalk delivery being a success.

But then…

In early March 2022, Tortoise pivoted again, vowing to focus entirely on mobile smart stores, which are essentially fancy vending machines placed on top of Tortoise’s delivery robots and located outside retailers. Now, Tortoise has moved from a hardware-as-a-service model to a take-rate scheme that gives it 10% of any sales made from its card payment-enabled bots, whether it’s a box of pastries from a bakery or brand new headphones from an electronics store.

Shevelenko, who served as Uber’s director of business development and was behind its acquisition of Jump bikes, says these pivots are just the beauty of a startup that’s responsive to market changes. The founder has advised or been on the board of a number of mobility and tech companies, including Skip, Superpedestrian, Codi, Payfare, Skyryse, SpotHero and Cargo Systems.

While Tortoise is his first time starting a company, Shevelenko is well versed in the factors that can cause a startup to win and lose.

We sat down with Shevelenko to talk about everything from Tier’s acquisition of Spin and the future of micromobility, how to own changing business directions, the difficulties in sidewalk robot delivery and the agility of startups.

The following interview, part of an ongoing series with founders who are building transportation companies, has been edited for length and clarity.

TC: At Uber, you were behind a lot of new mobility segments and the acquisition of Jump bikes. What do you think is the value of companies having multiple pillars, instead of just doing one thing really well?

Dmitry Shevelenko: For Uber, as a consumer-centric company, it’s ultimately a strategy of capturing all your transportation spend. The ultimate end state here — and this is why I think they’re putting so much money behind this Uber One subscription — is transportation-as-a-subscription product.

Ultimately, the way to win is to aggregate all the different ownership models so it’s shared, rented and owned. Dmitry Shevelenko

It’s not really effective for Uber and Lyft to try to win your business one trip at a time by offering you special incentives. If people are constantly switching back and forth between Uber and Lyft, they both lose. So the way to win is not by competing on a per-trip basis, but almost on an annual basis. How can you lock somebody in to be yours for a year? I think the essential nature of that consumer lock-in means you need to have more than just rideshare, right?

I think in rideshare, bundling is essential, because rideshare will have ups and downs. But the demand for transportation is constant. So if you have multiple modes, you’re always going to be doing well.

Tortoise’s original idea of repositioning scooters didn’t pan out in part because of the pandemic, but do you think it’s still a good idea?

Oh, absolutely. It’s just purely a function of sequencing and relative prioritization. The only reason delivery got so good, and there’s so much demand for it is because of COVID, too, right? It’s not only shared scooters that became bad.

Robotics founders: Build your pitch deck around problem-solving, not technology

In robotics, the remarkable often feels at odds with the practical. The Cassie robot captured the internet’s imagination (ours included) when it debuted in 2017 through a series of Oregon State University YouTube videos. It was one of the most exciting examples of robotics engineering since Boston Dynamics first made the scene.

Commercial applications, however, are a different conversation entirely. In a world of purpose-built systems, it’s not the first thing you see when you gaze upon the skinny legs of the ostrich-inspired bipedal ‘bot. When Agility Robotics first spun out of OSU’s College of Engineering, Cassie was being produced for research facilities. It’s a worthy mission, but not exactly a cash cow.

In a recent episode of TechCrunch Live, Agility’s co-founder and CTO, Jonathan Hurst, and Playground Global’s founding partner, Bruce Leak, joined us to discuss the robotic company’s journey from the lab to the commercial sector — and the role a good VC firm can play in that journey. The conversation spanned 30 minutes and includes a look at Agility Robotics’ early pitch deck. The deck and video are embedded below.

“If you’re building a company that’s building something that is really new and different, where are you going to hire engineers with experience with highly dynamic physical interaction, in the world, with force-sensitive behavior?” asks Hurst. “It’s just not common. Having students using the robots and a whole pipeline of people not only helps us, but it helps the whole infrastructure.”

From lab to launch

Playground Global, an early-stage investment firm based in Palo Alto, discovered the robot the way most of us did – watching cool videos online.

“We were surfing the internet like any good venture capital group, and we ran across the video that Agility released,” says Leak. “We were super impressed. This product, at some level, was just an incredible pair of legs. But it could walk for hours and even run across uneven terrain in a very practical way. Seeing something like that, which we thought might not even be possible, we knew we had to meet the Agility team.”

Agility’s seed/Series A pitch deck wasn’t focused on things like addressable market, and its insights into the robots’ practical commercial applications were cursory. What it did, however, was break down the startup’s impressive technologies. Hurst points to a tone shift between the presentation’s first slide, reading “Dynamic robots for human environments,” and its penultimate, “Made for work.”

Exploring the many faces of sidewalk delivery robots with Cartken’s Anjali Jindal Naik

Like many startup founders, Anjali Jindal Naik, co-founder and COO of autonomous sidewalk robot maker Cartken, was raised by entrepreneurs. Her parents owned a furniture store in North Carolina, and Naik spent much of middle school and high school helping out with managing warehouse deliveries, an experience that would later inform her current pursuits.

When she graduated from university, Naik’s father gave her some advice: Start your own business; don’t work for somebody else.

Naik followed her passion for Bollywood music and built her first company, Saavn, a successful distribution and streaming service for Indian and Bollywood music and content. At Saavn, Naik realized she liked to push the envelope with emerging tech and experiment with achieving product-market fit. Back in 2005, that meant working on ringtones for mobile phones, and even trying, and failing, to stream Indian concerts to mobile phones in the U.S.

Naik went on to handle operations and product for a number of companies, including, most notably, Google Express, a shopping service from Google that has since been swallowed by Google Shopping. It was there that she met the engineers over at the company’s Area 120 incubator for experimental products, Jake Stelman and Christian Bersch, who would later go on to become her co-founders at Cartken.

Stelman and Bersch worked on Bookbot, a sidewalk delivery robot that would deliver books from libraries. The project, and its pilot at Mountain View Library, was short-lived for business and political reasons rather than hardware or tech reasons – the robot reportedly operated quite well.

Sidewalks, to us, seem like the best way to get to an origin and an end destination. So that’s kind of where we’ve landed. Anjali Jindal Naik

That was in 2018. Cartken was formed the following year.

Since then, Cartken has started pilots with Reef Technology to bring food from Reef’s network of delivery-only kitchens to customers in Miami, with Erasmus University in Rotterdam to delivery convenience store items to students, and with Mitsubishi to provide indoor and curb-side delivery for Starbucks customers at a popular mall in Japan.

We sat down with Naik to talk about the benefits of graduating from a tech giant like Google, the rising demand in the robotic sidewalk delivery space, and how a baseline of strong tech can enable new form factors.

The following interview, part of an ongoing series with founders who are building transportation companies, has been edited for length and clarity.

TechCrunch: What’s your biggest takeaway as a startup that’s broken away from a larger parent company like Google?

Anjali Jindal Naik: When you do something under a larger umbrella, like Google, you do a lot of testing, trialing and prototyping. But I don’t know if it necessarily gives you the push that says, “Okay, let’s take this out to market and really move away from the safety net of doing this within a larger company.”

I think it’s nice to start a project in there. But if you really want to get the feeling of true entrepreneurship, going out on your own and maybe taking some of the knowledge and the tests that you’ve done, and creating something totally new outside of that umbrella is actually the best of both worlds. It gives you a little bit more confidence that what you’re putting out in the market has had some validation beforehand.

I don’t think we’ll ever escape the Google alum title. It is a core part of our story.

There’s a lot of debate in the industry about the best form factor for autonomous delivery. Why do you back sidewalk delivery?

I think being on the bike path or even on the road creates some barriers to entry. Sidewalks, to us, seem like the best way to get to an origin and an end destination. So that’s kind of where we’ve landed.

We have spent a lot of time working on our form factor to make sure that it’s not cumbersome and not a nuisance on the sidewalk to strollers, wheelchairs and others that need to share the sidewalk, but that there’s enough compartment storage to transfer whatever goods we need to transfer.

To cool down China’s overheated robotics industry, go back to the basics

It’s been a tumultuous few years, but China’s manufacturing industry is now on the rebound. Once an industry characterized by low-end manufacturing and intensive labor, it has transformed into a high-end manufacturing hub aided by technology.

Automation and robotics has the potential to modernize China’s manufacturing while improving labor efficiency and alleviating labor shortages. Predictably, companies and investors want to capitalize on this trend.

Robotics has been a hot sector for a while, but its popularity has shot up over the past couple of years. The sector recorded investments and financing of $6 billion in 2021, according to statistics from market research firms, and is expected to double in size in five years.

However, it’s unknown when these investments will provide a suitable return. Robotics is experiencing the biggest bubble in China’s venture capital industry, and is riddled with speculation and overvalued companies. Compared with similar investment bubbles over the last 10 years, this one is larger in scale, longer in duration, and could be more devastating than any before.

The price-to-earnings ratio is no longer applicable for many listed companies, and the market-to-sales ratio has also gone out the window. He Huang

However, the “bust” is entirely avoidable. Investors and companies need to go back to business basics and resist the industry’s typical impatience for exits on both sides of the negotiation table.

Understanding the market

With the influx of capital investment, we’re seeing a partial and cyclical overheating of the market in China. Many investors caught in this investment tide are replicating the software investment model, because many institutions that invested in Internet startups are also aggressively entering this field.

So what’s behind this surge? Everything from China’s government policy to the launch of the Science and Technology Innovation board, which has opened a convenient exit channel. Compounding the surge is the drive to upgrade China’s industrial structure.

It’s crucial, however, that investors do not apply software investment rules to industrial technology investments. For one, the investment to exit period is different. Investment in robotics and other industrial technologies is relatively long-term compared to internet companies. Internet companies can go public in three to five years after investment, but industrial technology firms are likely to take twice as long or more to go public.

The next healthcare revolution will have AI at its center

The global pandemic has heightened our understanding and sense of importance of our own health and the fragility of healthcare systems around the world. We’ve all come to realize how archaic many of our health processes are, and that, if we really want to, we can move at lightning speed. This is already leading to a massive acceleration in both the investment and application of artificial intelligence in the health and medical ecosystems.

Modern medicine in the 20th century benefited from unprec­edented scientific breakthroughs, resulting in improvements in every as­pect of healthcare. As a result, human life expectancy increased from 31 years in 1900 to 72 years in 2017. Today, I believe we are on the cusp of another healthcare revolution — one driven by artificial intelligence (AI). Advances in AI will usher in the era of modern medicine in truth.

Over the coming decades, we can expect medical diagnosis to evolve from an AI tool that provides analysis of options to an AI assistant that recommends treatments.

Digitization enables powerful AI

The healthcare sector is seeing massive digitization of everything from patient records and radiology data to wearable computing and multiomics. This will redefine healthcare as a data-driven industry, and when that happens, it will leverage the power of AI — its ability to continuously improve with more data.

When there is enough data, AI can do a much more accurate job of diagnosis and treatment than human doctors by absorbing and checking billions of cases and outcomes. AI can take into account everyone’s data to personalize treatment accordingly, or keep up with a massive number of new drugs, treatments and studies. Doing all of this well is beyond human capabilities.

AI-powered diagnosis

I anticipate diagnostic AI will surpass all but the best doctors in the next 20 years. Studies have shown that AI trained on sizable data can outperform physicians in several areas of medical diagnosis regarding brain tumors, eye disease, breast cancer, skin cancer and lung cancer. Further trials are needed, but as these technologies are deployed and more data is gathered, the AI stands to outclass doctors.

We will eventually see diagnostic AI for general practitioners, one disease at a time, to gradually cover all diagnoses. Over time, AI may become capable of acting as your general practitioner or family doctor.

5 fundraising imperatives for robotics startups

Early-stage robotics fundraising is accelerating, with funding coming from boutiques to deep-pocketed venture capital firms. For founders, getting their idea from concept to company, or developing a minimum viable product, is daunting enough, but seeking an initial fundraising round brings a complexity that can be especially challenging to manage.

So how do robotics startups best approach fundraising and secure the financing to propel their company to the next level? There are five key areas to keep in mind about fundraising for robotics startups that founders must learn and practice.

Understand the proper fit between your company’s scale and the fund’s scale

Too often, founders court venture capitalists without understanding that the company they are founding might not be the right fit for VCs. Venture capital firms generally, and ones that invest in robotics specifically, look to invest in startups that have clearly identified potential to scale exponentially.

They are not geared toward backing entrepreneurs looking for an exit under $100 million that will only realize a handful of multiples for the investor. VCs are more likely looking to fund on a much larger scale — think a $1-billion-plus exit valuation — and back a company with the potential to deliver at least a 10x return.

Venture capital firms generally, and ones that invest in robotics specifically, look to invest in startups that have clearly identified potential to scale exponentially.

Usually, robotics companies are capital-intensive and require a robust revenue model compared to pure software startups, and this is not for every VC. In fact, venture capital is likened to “rocket fuel” that is dangerous if put into a car but perfect for a rocket ready to shoot for escape velocity. Smaller-scale ventures often do not interest VCs but might be perfect for angel investors.

Bottom line: Do your homework, manage expectations, and seek funding from investors working at a scale commensurate with your idea and comfortable with the unique needs of robotics companies.

Consider capital sources that fit different companies and startups at different stages of its growth

Now is a great time for starting a company, in part because there have never been more sources of financing available. Angel investors and venture capitalists are just a portion of what is available.

There is growing opportunity, especially for robotics and AI startups, in nondilutive capital, including from U.S. government sources such as Department of Energy and Department of Defense grants. There are loaded/nondilutive funding streams, such as convertible debt, available from financial institutions and angels.

Special purpose acquisition companies, or SPACs, especially for hardware and robotics companies, have become popular in recent years. Some of these might be a better fit for your company at the current (or future) stage of your organizational growth cycle.

But some sophistication is warranted. Ask yourself what constraints or potential downsides come with the specific funding model you are considering/pursuing. Government grants, for instance, might drive the pace of development or push you toward certain customer-facing directions in ways that could be ill-suited to your company.

Refraction AI’s Matthew Johnson-Roberson on finding the middle path to robotic delivery

Refraction AI calls itself the Goldilocks of robotic delivery. The Ann Arbor-based company, which recently raised a $4.2 million seed round and expanded operations to Austin, was founded by two University of Michigan professors who think delivery via full-size autonomous vehicles (AV) is not nearly as close as many promise and sidewalk delivery comes with too many hassles and not enough payoff. Their “just right” solution? Find a middle path, or rather, a bike path.

The company’s REV-1 robot, which co-founder and CTO Matthew Johnson-Roberson debuted on the TechCrunch Sessions: Mobility stage in 2019, was built on a foundation of a bicycle. At about 4 feet tall and 32 inches wide, the three-wheeled vehicle can travel at up to 15 miles per hour, which means it can stop quickly to avoid obstacles while still being faster than a human.

The intermediate speed also means that the REV-1 doesn’t need to see as far ahead as a full-size AV, which allows it to function well on radars, sensors and cameras instead of requiring expensive lidar, according to the company.

Johnson-Roberson has spent nearly 20 years in academic robotics. Universities are home to many of the advances in field robotics, but the average person doesn’t see many such applications everyday when they look out their window. This desire to make something that is useful to the general public has been a huge motivator for the academic-turned-founder.

The following interview, part of an ongoing series with founders who are building transportation companies, has been edited for length and clarity. 

TechCrunch: You unveiled Refraction AI on the TechCrunch stage two years ago. How has it evolved since?

Matthew Johnson-Roberson: It’s been a really exciting ride. At that time, we had one vehicle — the one that we rolled out on stage — and now we have 25 vehicles in Ann Arbor and Austin, which we just announced. So things have changed quite a bit in the intervening years. We had already predicted a lot of changes around food delivery, specifically, and lots of those were accelerated by the pandemic.

Deep Science: Robots, meet world

Research papers come out far too frequently for anyone to read them all. That’s especially true in the field of machine learning, which now affects (and produces papers in) practically every industry and company. This column aims to collect some of the most relevant recent discoveries and papers — particularly in, but not limited to, artificial intelligence — and explain why they matter.

This edition, we have a lot of items concerned with the interface between AI or robotics and the real world. Of course most applications of this type of technology have real-world applications, but specifically this research is about the inevitable difficulties that occur due to limitations on either side of the real-virtual divide.

One issue that constantly comes up in robotics is how slow things actually go in the real world. Naturally some robots trained on certain tasks can do them with superhuman speed and agility, but for most that’s not the case. They need to check their observations against their virtual model of the world so frequently that tasks like picking up an item and putting it down can take minutes.

What’s especially frustrating about this is that the real world is the best place to train robots, since ultimately they’ll be operating in it. One approach to addressing this is by increasing the value of every hour of real-world testing you do, which is the goal of this project over at Google.

In a rather technical blog post the team describes the challenge of using and integrating data from multiple robots learning and performing multiple tasks. It’s complicated, but they talk about creating a unified process for assigning and evaluating tasks, and adjusting future assignments and evaluations based on that. More intuitively, they create a process by which success at task A improves the robots’ ability to do task B, even if they’re different.

Humans do it — knowing how to throw a ball well gives you a head start on throwing a dart, for instance. Making the most of valuable real-world training is important, and this shows there’s lots more optimization to do there.

Another approach is to improve the quality of simulations so they’re closer to what a robot will encounter when it takes its knowledge to the real world. That’s the goal of the Allen Institute for AI’s THOR training environment and its newest denizen, ManipulaTHOR.

Animated image of a robot navigating a virtual environment and moving items around.

Image Credits: Allen Institute

Simulators like THOR provide an analogue to the real world where an AI can learn basic knowledge like how to navigate a room to find a specific object — a surprisingly difficult task! Simulators balance the need for realism with the computational cost of providing it, and the result is a system where a robot agent can spend thousands of virtual “hours” trying things over and over with no need to plug them in, oil their joints and so on.