Venture Capital & Mobile Robotics: Bad Investment?


Posted on December 22, 2020 by John Ripple


Warehouse automation, including automatic guided vehicles (AGVs) and autonomous mobile robots (AMRs) is a hot market for VC investment. In 2020, the top 15 companies in terms of VC funding received investments totaling $888 million USD. These companies represent either the next dot.com bubble or an entry point to the Fourth Industrial Revolution that you don’t want to miss.

All companies on the 2020 VC funded list build some version of “mobile robotics” used for warehousing and distribution applications. Each supplier has their unique take on equipment and application, but at the end of the day we’re talking about automated equipment to move cartons, totes, and boxes around a warehouse. This article is not intended to be a comprehensive “tech-stack” review. The purpose of this article is to explore whether smart people are making dumb investments.

The top VC backed companies of 2020 for mobile robotics are:

Company
VC Funding 2020
Total Cumulative
Berkshire Grey
$263 mil.
$266 mil.
GreyOrange
$100 mil.
$270 mil.
Exotec
$90 mil.
$112 mil.
Seegrid
$52 mil.
$150 mil.
Geek+
$50 mil.
$390 mil.
Attabotics
$50 mil.
$83 mil.
Vecna
$50 mil.
$63.5 mil.
Fetch
$46 mil.
$94 mil.
Locus
$40 mil.
$105 mil.
Covariant
$40 mil.
$67 mil.
Wayz
$36 mil.
$36 mil.
OTTO Motors
$29 mil.
$83 mil.
Third Wave
$15 mil.
$15 mil.
Youibot
$15 mil.
$15 mil.
CAJA Robotics
$12 mil.
$26 mil.
$888 mil.
$1,775 mil.

The warehouse automation market is growing rapidly. Global retailers such as Amazon, Walmart, and Alibaba have embraced warehouse automation as a core strategy propelling business growth. Companies that fail to embrace automation will become less competitive and lose market share.

This article will evaluate the mobile robotics segment of the warehouse automation industry and these companies based on:

  1. Market opportunity – what size is the pie?
  2. Financial valuations – what does it mean?
  3. What kind of sales volumes support the valuations?
  4. Projected sales based on investment – is it even possible?
  5. Pricing and Made in USA vs. Made in China

We’ll look at these companies mostly as a group, though we’ll call a few out by name here and there. Some of these companies will go on to be highly successful. Some will fail. This article isn’t about picking winners or losers; it’s about how the game is played. Buckle up. It’s time to run the numbers!

1. Market opportunity – what size is the pie?

It is difficult to get a good handle on the size of the market. A widely publicized market research study that gets shared around the industry indicates the following with respect to the market for warehouse automation:

LogisticsIQ™’s latest market research study “Warehouse Automation Market By Technology (AGV/AMR, AS/RS, Conveyors, Sortation, Order Picking, Automatic Identification and Data Capture, Palletizing & Depalletizing, Overhead Systems, MRO Services and WMS/WES/WCS), By Industry (E-commerce, General Merchandise, Grocery, Apparel, Food & Beverage, Pharma, 3PL), By Functions (Inbound, Picking, Outbound), By Geography – Global Forecast to 2025“, estimates that the Global Warehouse Automation Market will grow more than 2x from $13 Billion in 2018 to $27 billion by 2025, at a CAGR of 11.7% between 2019 and 2025.

Several websites re-publish the LogisticsIQ market research. For example, Statista.com and Businesswire.com both publish $13 billion as the size of the warehouse automation market in 2018. The LogisticIQ study has been shared around the industry (including by some of the Top 15 VC backed companies) as evidence of rapid growth in warehouse automation.

Modern Materials Handling publishes an annual list of the Top 20 System Suppliers for warehouse automation. The 2019 Top 20 list indicated that the reported 2018 revenue from the Top 20 suppliers in the warehouse automation industry was $21.5 billion. In fact, the top 7 warehouse automation suppliers alone exceeded the $13 billion in sales for 2018 referenced by LogisticsIQ.

Clearly something amiss. While the actual numbers are unclear, it is likely that the market opportunity is 2X or more the widely reported market figures. Great news. But a good reminder to cross-reference sources.

What is clear is that warehouse automation is growing and has a long runway ahead. There are many warehouses and distribution centers across North America and around the world with little in the way of mechanized automation. Increased adoption is the challenge and the opportunity.

2. Financial valuations – what does it mean?

VC investing is risky business. There are 3 common rules-of-thumb:

Rule 1: 1/3 of VC investments will fail, 1/3 will provide marginal returns or just return the original investment, the remaining 1/3 provide the juicy portfolio returns. Everyone loves a unicorn.

Rule 2: Rule of Two. This rule of thumb seems like it was conjured up by someone with years of experience in the automated material handling industry. Take a startup’s projections and multiply by 2, multiply by 2, and divide by 2. That is, businesses often require 2x projected capital needs, take 2x as long to reach the investment goal, and attain 1/2x of projected revenue. Ouch!

Rule 3: Rule of 10X. Investments should return 10X over 10 years. Every company on our Top 15 list is aiming for the moon. They wouldn’t have been VC backed otherwise. Go big or go home!

The Top 15 VC-funded mobile automation companies raised $888 million in 2020 alone. These figures do NOT include mergers and acquisitions. They do NOT include R&D investment by established industry players. This does not count private investment. This is new VC investment in emerging suppliers.

The same group of VC backed companies that raised $888 million in 2020 have raised $1.775 billion dollars cumulatively amongst themselves over the last few years. Let’s apply the Rule of 10. For simplicity we’ll start the clock now and say the group needs to return 10X the investment by 2030. In reality, the clock has been ticking for many of these companies for a few years now. That would indicate the group should be worth $17.75 billion in 2030. Mind boggling valuations for a group of startups in the world of warehouse automation.

Before speculating about P/E multiples and valuations, let’s take a look at history. In 2016, industrial conglomerate KION Group acquired warehouse automation industry giant, Dematic. 2015 revenue for Dematic was $1.8 billion with an EBIT of $166 million. The purchase price was $2.1 billion, resulting in a P/E multiple of 13. Dematic was doing well and was forecasted to continue doing well. It was not a company in distress. Another metric to keep in mind: in 2016, Dematic had 6,000 employees, resulting in revenue per employee of $300,000.

One more historical data point, then we’ll move along to our group of VC backed startups. In 2016, Honeywell acquired Intelligrated for $1.5 billion. The purchase price translated to approximately 12 times estimated 2016 earnings before interest, taxes, depreciation and amortization. At the time they had 3,100 employees with a revenue per employee of $290,000.

Based on investments already made in Berkshire Grey and GreyOrange, applying a 10X multiple, each of those companies individually should be worth more than Dematic was in 2016 in the next few years.

Under the premise that these VC backed companies should be worth $17.75 billion in 2030, thereby qualifying themselves as “good investments in the VC world”, what does that mean? Well, a company is only worth what someone will pay for it, right? The S&P 500 has a 12-month forward P/E ratio of 19 and a 10-year average of 15. We should assume that our hot-tech VC starlings are all above average and assume that a P/E ratio of 25 is not unreasonable. Perhaps a P/E multiple of 25 is low-balling a high-growth company, but we need to start somewhere. Feet grounded in reality and 2X the industry veterans is as good a spot to start as any.

3. What kind of sales volumes support the valuations?

While we are grounded in reality for financial valuations, a P/E multiple of 25 has the effect of requiring less than industry standard revenue to achieve greater than industry standard valuations.

Using a P/E ratio of 25 on a valuation of $17.75 billion would indicate these top 15 VC-backed mobile automation companies are churning out earnings of $710 million per year in 2030. Sure, that is a big number, but what does it mean in practical terms to an AGV/AMR supplier? These mobile automation companies are producing robots of various types that have a sell price in the range of $30,000 each with about $3,000 in profit per each robot sold. Some are more, some are less, but as a general industry metric, not too far off.

That means as a group, in the year 2030, they should be cranking out 237,000 robots per year! Is that feasible? We’ll get to that. But, based on money already invested by the VC backers, assuming no further investments needed (Rule of 2 need not apply!), assuming a reasonable rate of return for the VC backer, including a reasonable valuation for a future buyer (they must be sold for the VC backer to realize those returns), those are the numbers that lead to the required valuations.

The top 5 in terms of funding have some serious work to do: Geek+ has raised $390 million and by extension should be cranking out 52,000 robots per year in 2030. GreyOrange is responsible for 36,000 per year. Berkshire Grey needs to deliver 35,467 per year. Seegrid needs to see 20,000 per year out the door. Exotec only has 14,933 per year to their name. Phew! If you’re feeling like Dematic or Intelligrated was the better benchmark for earnings potential, then double those numbers.

We arrived at the number of robots based on earnings required to justify a P/E multiple. Confusing, yes. But the purpose of the company is to make money for the investors, not just make a bunch of robots! If we take the number of robots they’d need to produce to support their valuation (which is driven by investment – ooh, it’s a vicious cycle), multiply by $30,000 per robot, you arrive at a total sales volume of $7.1 billion for the group based on the aggressive approach, or double that for the old-school Dematic approach.

All this talk of P/E and valuation is a long way of saying that our VC backed group is expected to be worth a lot of money with only half the revenue dollars that is typical in the industry. As we’ll soon explore, getting to even ½ those revenue dollars is hard work!

4. Projected sales based on investment – is it even possible?

Warren Buffet offered sage advice: don’t invest in businesses you don’t understand. These numbers all sound great, but at the end of the day you need customers, projects, and people to design, sell, and implement those projects.

Who knows what the year 2030 will bring, but we can take a look at 2020 for a benchmark. In 2020, a large AMR-based robotics project included 100 robots of various types (realistically, most projects are sub-50 robots, but we’re aiming on the side of generosity). Middle of pack contributor, Berkshire Grey needs to complete some 350+ projects per year with 100 robots per project. If projects only have 50 robots, then some 700+ projects per year.

That number of large-scale projects is no small feat. The sales funnel for warehouse automation projects with AGV/AMR/robotics technology for multi-million-dollar projects (the type with 100+ vehicles) goes something like this:

Each successful salesperson will talk with 50-60 different customers in the course of a year that express interest and have a potential project. Of those 60 potentials, about 30 pass the initial sniff test for qualification and warrant further conversations. Of those 30 customers that warranted more than 1 or 2 conversations, about 10-12 are worth going after with real projects and real potential. Each one of those 10-12 require the salesperson to get in the car or on an airplane to go visit the customer, see the application, and make the pitch. About 8 of those opportunities get serious and require multiple trips (usually 4 trips to seal the deal). Per year, about 6 projects per salesperson turn into orders. Typical sales cycle from 1st discussion to order is 18 months.

Remember: we’re talking about multi-million-dollar orders here. The types of projects required to move the needle on sales volumes. They take time to develop a solution, client trust, financial justification, approval, and permission to proceed. These are systems that are deeply integrated into a clients’ operation. They can make or break a business. It takes time and expertise to sell. There is no magic pill. There is no solution so spectacular that it sells itself.

Knocking down hundreds of projects per in 2030 per each company will be a challenge. Not every project will include 100+ robots. Not every salesperson will enjoy high levels of success. These companies not only need to develop their technology, but they’ll also need to compete against established industry players, and win new business, consistently. They will need to scale their sales organizations with individuals that understand how to sell complex engineered systems and equipment to sophisticated corporate buyers.

Based on valuations, Berkshire Grey will need to produce 35,000+ robots of various types with more than 3,500 employees in 2030. As of December 2020, Berkshire Grey had 223 employees on LinkedIn. They’ll need to add more than 3,000 employees over the next few years to hit their valuation and growth targets. Roughly 1 person every business day between now and 2030. If their sales revenue per employee resembles Dematic and Intelligrated, they will need 7,000+ employees.

5. Pricing and Made in USA vs. Made in China.

Before we discuss country of origin, this is a good place to address pricing, especially as it relates to early “sales success” enjoyed by some entrants to the market. Some VC backed robotics players have gained market entry and initial sales volume by subsidizing pricing. The prices offered to customers are not based on costs + profit, they are based on “winning the project” (a.k.a. VC investment subsidizes price). Of course, that is not a surprise and perhaps even necessary in the early phases. However, it is not a sustainable business practice. It falsely inflates sales volumes and doesn’t predict future success. If the customer had to pay the true price for the solution, would they have purchased that solution?

Pricing and the China-factor are inextricably linked. Whether or not the China-factor should be included in the same discussion as price-subsidies is up for debate. However, we can consider a few metrics that are known facts and steer clear of politics.

Some of the major producers of mobile robots in China work on “996” schedule. That means they work from 9am to 9pm, 6 days per week. They are working 50% more hours than a typical American or European employee. That level of effort does a few things: it expedites the R&D process and they complete 1.5X more work for the same number of people. Their employees are highly educated and motivated. Compared to American’s and European’s, they have lower levels of compensation. The North American and European based companies are going to have to hustle to keep up with the 996 crowd.

Suppliers from China are selling their mobile robots at price levels far below the North American suppliers. For example, a typical AMR from a North American supplier that handles 2,500lbs loads sells for $65,000 to $80,000. A similar AMR from a Chinese supplier is sub-$30,000. A smaller form factor AMR that handles between 100lbs to 600lbs sells for $30,000 to $40,000 from North American suppliers. Comparable units from China are sub-$20,000. Big differences.

It is still early to see how it plays out in the North American market with mobile robots “Made in the USA” vs. “Made Overseas”. In 2018, there were very few entrants in the USA market from Asia offering mobile robotics. System pricing was wildly different from supplier to basis based on perceived risk and complexity, customer demands, and whether equipment was subsidized or not. The years 2019 and 2020 saw an influx of new suppliers to the North American and European markets from Asia.

It is likely some of the “Made in the USA” suppliers will struggle with competitiveness (and increasing market share) vs. “Made Overseas” as the solutions and equipment become more widely adopted. Service and implementation challenges notwithstanding.

6. Robots as a Service (RaaS) and the software play.

Some of the VC backed players offer Robots as a Service (RaaS). These are mechanical machines, not pure software plays. The robots have a circa 5-year lifespan in operation. They break down, wear out, and become obsolete. At the end of 5 years, their residual value is zero. Some robots may go on to have a longer service life; but fully amortized. RaaS is really a lease, maintenance included, but with a trendy name.

Some suppliers offer “elastic fleet sizes”; they’ll provide additional equipment during peak periods and then you don’t pay for that equipment during off-peak periods. That’s great, except for the fact that nearly everyone in a given market has the SAME peak period. This means 1 of 3 things: One, the additional robots won’t really be available when the customer needs them – not good for customer. Two, the customer is actually paying for the total fleet size they need for the peak period anyhow – nobody is fooled there. Three, the VC investor is subsidizing a fleet of vehicles in stock – someday somebody will need to pay the piper.

Many of the robot suppliers claim they are software companies and therefore should justify the P/E multiples and valuations of pure software plays. That is a good thought, except that nobody purchases the robot without the software or vica versa. The equipment and software are inextricably linked. If the hardware goes away, the software (revenue) does too.

7. Competition: It’s a small world out there.

We’ll leave the subject of China and North American pricing aside and look at the competitive landscape and market opportunity. Over the last 25 years, a constant theme is that hyper-growth and radical acceptance of warehouse automation is just around the corner. Give the industry credit for sustained optimism! Has the industry grown? Absolutely. Year over year industry growth of 10% is a fair figure. The industry is well ahead of inflation, but well shy of producing 10X growth in 10 years without innovation and increasing levels of robotics adoption.

The warehouse automation industry has a few challenges:

  1. Large projects are company and career “make or break” events. Asking an organization to invest in a large-scale automation project is asking management to put their careers on the line. Trendy tech’ and risky designs are a tough sell.
  2. Multi-million-dollar investments take time. Projects take time to develop the business case, obtain management approval, and navigate the bureaucracies of the companies that need large-scale projects.
  3. Established veterans of the warehouse automation industry have been working with the largest customers for years and have developed bespoke solutions (and trust) that are proven with their customers.
  4. Automation equipment, solutions, and hardware are frequently not the largest costs related to warehousing and distribution investment. The building, land, and infrastructure is usually many times the cost of automation. Companies are unwilling to risk that a $50 million distribution center doesn’t work as intended because they selected an unproven automation solution.
  5. Winning new projects builds on a reputation of successfully completed projects. Because warehouse automation projects have a long sales cycles and long implementation cycle, building traction in the market takes time. VC investors may need be patient.

The warehouse automation industry does offer some exciting opportunities:

  1. The industry is desperate to reduce labor required to pick, pack, and ship orders. Not because they want to lay people off, but because they can’t find enough people. Personnel turnover in warehousing operations is frequently 100% or more per year. If a warehouse has 250 people, they likely need to recruit and train 300 people per year.
  2. By 2026, the labor force participation rate is projected to drop an additional 2%, due largely to an aging population. A shrinking workforce is inevitable. Many sectors in the economy already face hiring challenges, and the competition for employees is likely to get even more intense.
  3. Warehouse automation is a competitive advantage for those that embrace it as a strategic advantage. The ability to deliver products to consumers quickly and at low costs is a competitive advantage enabled by warehouse automation technologies, including mobile automation.

Conclusion

The VC investing rules of thumb are likely to be alive and well for investments in mobile robotics. For some of these suppliers, it will take 2X longer and 2X more investment to achieve ½ their revenue goals. Some of the companies on the list will fail; perhaps not in an absolute sense, but in terms of providing a return to their VC investors. The main challenges will be:

  1. Customer adoption of emerging technology at scale.
  2. Finding and retaining people with experience to design and sell complex solutions.
  3. Designing and supporting robust equipment that survives 24x7 operation.
  4. Retaining competitive advantage in an increasing global marketplace.
  5. Satisfying growth targets in an industry with long sales cycles.

The winners and losers are employees and customers. They will reap the rewards of success and feel the pain of failure. Warehouse automation is a good business with good growth. The question to be answered is if VC investors have over-invested in some companies such that return of capital is at risk. The answer to that is most certainly yes. Dumb money? No. Investing in a business they don’t understand? Perhaps.

Categories