Space to Watch: Robotics for Food Processing

Written by Bogdan Cristei of Shack15 Ventures and Tim Yang of Tyche Partners

Bogdan Cristei

13 Dec 2022 — 12 min read

Written by Bogdan Cristei of Shack15 Ventures and Tim Yang of Tyche Partners

The current labor crunch is incredibly acute, with plants operating at ~60% capacity. Labor is gone and not coming back, so the industry has to think deeply about automation. There is a clear need for robotics, however, robotics companies need to hit adjacent markets and serve more than one industry to survive. When looking at companies in this space, investors need to look for ability to interact and operate with a wide variety of products and line configurations, as well as access to proprietary data. This article goes over an overview of the industry, current state of automation, competitive landscape, industry gotchas and what to look for when evaluating investments in robotics for food processing.

Table of Contents:

Automation
Robotics
Safety
Labor
Unions
Plants, Facilities & Logistics
Chickens vs Pork vs Cattle
Business Model Gotchas
Business Model Options
What to Look For in a Potential Investment
Why Now?
Risks
Market Size
Barriers to Entry
Competitive Landscape
Research Flags
Industry Players
Industry Stats
Current Technologies
References

1. Automation

Folks are looking at automating low hanging fruit vs critical path items at the moment— things like packaging, labeling and other low skill repetitive actions
Plant owners want to automate critical points on the manufacturing line, but they are hard to automate; for example when considering cattle, a cow can range between 600 lbs –1000 lbs; it is difficult to build a “one size fits all” robotic process, and skilled laborers are needed to deconstruct large animals of various sizes
Hard to gain traction in the beginning for automation — customer has low tolerance in supply chain for pauses, given the high turnover nature of the business
Some folks looked at automated split saws, but hard to implement — a lot of equipment is big and doesn’t fit into footprint, and there is a need to rebuild fabrication floors

2. Robotics

Robotics companies active in the space need to hit adjacent markets due to low TAM; they need to be able to serve more than one industry
Throughput is very important - food spoils and expires; if you don’t have human capital or resources, it makes sense to sell a bigger piece of an animal at a loss and keep the inventory moving vs waiting to cut up into smaller pieces and increase margins

3. Safety

Covid hit industry hard. Covid mitigations are real, CEOs are called before Congress, and they are getting serious about automation. For example, Smithfield wants to automate 500 positions per year for next 3 years. Every other company is similar, Tyson has an automation division now as well.
Safety-wise, the industry suffers 4 amputations per year. Food processing is described as “most dangerous menial job you can have” per recent articles; 4 out of 100 folks get off the line via injuries.

4. Labor

To operate at peak production capacity meat processors must employ a large number of facility workers. These workers complete tasks on various lines including slicing, packaging, and stamping. The repetitive nature of these tasks presents a strong opportunity for automation.
The labor crunch is incredibly acute, plants are operating at 60%-70% of capacity. Everyone has labor shortages. Facilities are 40 degrees and freezing, they smell bad, and folks would rather work in a more comfortable Amazon warehouse at higher wages.
Labor is gone and not combing back, so the industry has to take the robots and think deeply about automation. There are a few supporting US Government programs for small organisations, but financial help is already distributed and gone.
Labor fluctuations mean you need to decide what to do with labor since it is limited; you may be able to get some tasks done but not others
There are some jobs that are most important / without which the plant doesn’t run — example is splitting a carcass, gutter, deboning, etc… maybe check those and figure out solutions for those if you are starting a new company — critical components — not just at beginning of line, but across the entire thing
You don’t run if you don’t harvest — you need to make sure you harvest; deboning and such is not happening sometimes, product going in places shouldn’t be going; you need to get the animals dead and get them into whatever format you can sell them — some folks are incapable of capturing market opportunity due to limitations

5. Unions

In terms of unions, there is a generally a 50/50 unionized environment for facilities; robotics is not too great a concern given labor is going down — in the next few years nobody will lose jobs. All-in cost per employee is 75k. They tried higher wages, but still can’t find enough people.

6. Plants, Facilities & Logistics

How often do line revamps happen? One line undergoing revamp at any given time usually + tend to happen every couple of years. No need to wait for sales cycle. Will automate quickly after they see it working on one line. Plants go down due to cleaning, mostly germs from humans; w/ robots you can keep online more often.
Usually can’t maintain highest value “upgrades” — you have to keep facility moving with perishable goods; you can only hold 3 days of production — incremental revenue stuff often gets deprioritized
Most of the facilities are 30–60 years old, they are built for commodity businesses, they have been adapted to do more jobs not originally made to do; sometimes there is a need to scale back value add to keep things moving

7. Chickens vs Pork vs Cattle

The Chicken + pork industry is mostly the same everywhere; in the beef world you have different farmer feeders, genetics are different, equipment is different — all animals are different — a cow can vary between 600 - 1000 pounds for example — this is one of the excuses why technology doesn’t work there.
Base rate for the industry is 20–22$/hour; top is 30–32$/hour

8. Business Model Gotchas

SaaS most likely will not work; industry is not keen on subscriptions and memberships; once they commit, they want long term implementation vs subscription
Industry gross margins are low, hovering around 10%–15%
Industry CapEx per large player (eg Tyson) is ~1.5B per year and includes buildings, factories, supply chains — if we assume 10%-15% of that is for equipment we get to ~$100M-$200M per year; There are ~6 players in totay, so call it a $500M–600M TAM, which is fairly low
Yearly industry inventory is about $4B, COGS is about $44B; they turn ten times per year, meaning less than 1x/mo inventory turnover; they get the meat and in one month they need ship the product
Beef purchased at Whole Foods can last about one month max on the shelf; this is a tough business
Irregularly shaped inputs
Harsh and unattractive environment for workers and robotics engineers

9. Business Model Options

Upfront payment with monthly service, pay for robots upfront, allowing for amortization and depreciation of the asset, and service contracts for software
Monthly payments with no upfront payment, client doesn’t pay upfront for robots, instead pays monthly service fee and leases robots for a number of years
Full upfront payment no monthly payments, client pays for both service and robot upfront allowing for amortization of both services under one lump sum or as annual payment

10. What to Look For in a Potential Investment

Ability to interact and operate with a wide variety of different products and line configurations.
Force sensing sensors
Small footprint, only taking up as much space as a single employee on the line
Access to proprietary data from the industry; good quality data, directly from potential users
Culture and communicate here is important to consider; a green team out of college would be at a disadvantage trying to solve problems in this space, simply because they need to learn how to communicate with an industry which is conservative and relationship based

11. Why Now?

Macro conditions in the labor market are beginning to deteriorate from an already abysmal condition post COVID
Labor shortages: demand for simple repeatable task workers is through the roof, yet people do not feel motivated to work unless pay is much higher than a couple years ago. This opens the door for automated solutions that can reduce expensive payroll for somewhat menial jobs.
Increasing labor expenses, high concentration of labor-intensive processes
The industry want to automate (e.g. Tyson plans to spend more than $1 billion automating their facilities over the next four years [7])
Supply chain issues due to labor constraints
High Turnover — Industry has around 160% Churn [3]
Robust customer base
Slow rate of innovation
Large and defensible market
Few Competitors

12. Risks

Low skilled workers may need to be trained to use new technologies
Deployment & learning curve time may become bottle neck
Competitors in the space are well positioned to fast follow
IP protection is low when hardware is off the shelf
There are not too many food industry players in total; this can be good or bad, especially as it relates to IPO timing; If already working with all the players, may be hard to IPO as growth story may not land well

13. Market Size

Meat Processing Automation Market : $7B [5]
Food Automation: $28B [4]
Protein processing industry $227.6B [4]
Industrial Automation: $265B [6]

14. Barriers to Entry

Regulations
Food safety certifications
Cleaning Requirements

15. Competitive Landscape

Altitude
Website www.altitudeai.ai

Altitude AI Is The Brains Of Industrial Automation; Our software platform, Altitude OS, enables industrial robots to complete advanced never before seen tasks. Services include “Pick and Place” - our food safe pick and place solutions are capable of interacting with a wide variety of packaged or unpackaged products with ease. Thanks to our best-in-class computer vision we are able identify and interact with products in realtime which results in accurate and reliable placement of your products, as well as “Cutting and Trimming” — our hygienic cutting and trimming solutions are capable of interacting with a wide variety of existing tools to complete complex tasks. Through pairing our best-in-class computer vision with 6 dimension force sensing we are able identify and interact with products in realtime which results in surgical level precision and significantly improved yield.

Bright Machines
Website www.brightmachines.com

Provider of an automation software designed to assist businesses to meet the growing demands of manufacturing. The company’s software leverages artificial intelligence, machine learning, and production data to develop robots that are sensor-rich, have computer vision, and are adaptable, enabling clients to have improved yields at lower costs.

Year Founded 2018
HQ Location San Francisco, CA
Total Raised 315.45
Last Financing Valuation 871.45
Last Financing Deal Type 2 Series B
Last Known Valuation 871.45
Last Known Valuation Date 20-Apr-22
Last Financing Date 20-Apr-22
Last Financing Size 71.45

Competitors
Sight Machine, KONUX, C3.ai

Active Investors
Balyasny Asset Management, BMW i Ventures, Eclipse Ventures, EDBI, Flex, Geodesic Capital, Lux Capital

Chef
Website www.chefrobotics.ai

Developer of a food technology designed to reduce the cost of fresh food using robotics. The company’s technology is designed with modular hardware and reprogrammable software that uses computer vision technology and robotics, enabling restaurants to manage peak hours during lunch and dinner and prevent human errors during the preparation of food.

Year Founded 2019
HQ Location San Francisco, CA
Total Raised 10.2
Last Financing Deal Type Debt — General
Last Financing Deal Type 2
Last Known Valuation 20.5
Last Known Valuation Date 1-Jan-21
Last Financing Date 3-Jun-22

Active Investors
Adrian Macneil, Alex Feerst, Asymmetry Ventures, Bloomberg Beta, Bold Capital Partners, Construct Capital (Washington DC), Dave Kashen, Gaingels, Gokul Rajaram, Humbition, Jim Ostrowski, Kleiner Perkins, Pratik Agarwal, Promus Ventures, Red & Blue Ventures, Sal Candido, Schox Ventures, Soma Capital, Stewart Alsop, Tau Ventures, Vikas Enti

Covariant
Website www.covariant.ai

Developer of AI-based robots and software designed to create a roadmap and deploy robotics across operations. The company’s software engages in deep imitation learning, deep reinforcement learning, and meta-learning, enabling users to program a robot by simply donning a VR headset and guiding a robot through a task.

Year Founded 2017
HQ Location Berkeley, CA
Total Raised 147
Last Financing Valuation 530
Last Financing Deal Type 2 Series C
Last Known Valuation 530
Last Known Valuation Date 27-Jul-21
Last Financing Date 27-Jul-21
Last Financing Size 80

Active Investors
11.2 Capital, A&E Investments, A.Capital Ventures, Amplify Partners, Anca Dragan, Baidu Ventures, Canada Pension Plan Investment Board, Daniela Rus, Fei-Fei Li, Frees Fund, Geoffrey Hinton, Index Ventures, Industry Ventures, Jeff Dean, Kurt Keutzer, Lux Capital, Michael Jordan, Radical Ventures, Raquel Urtasun, Samsung NEXT Ventures, Scott Shenker, Silvio Savarese, SV Angel, Temasek Holdings, Trevor Darrell, Volodymyr Mnih, Xiaodong Song, Yann LeCun

Delkor Systems Inc.
Website www.delkorsystems.com

The company primarily operates in the Commercial Products industry. Delkor Systems Inc. was founded in 1973 and is headquartered in Saint Paul, MN.

Year Founded 1973
HQ Location Saint Paul, MN

Intrinsic
Website www.intrinsic.ai

Intrinsic is a robotics software and AI company at Alphabet. Our mission is to make industrial robotics accessible and usable for millions more businesses, entrepreneurs, and developers. We are a multidisciplinary team of engineers, roboticists, designers, and technologists who are passionate about unlocking the creative and economic potential of industrial robotics.

Year Founded 2021

Scott Technology (NZE: SCT)
Website www.scottautomation.com

Scott Technology Ltd is a robotics and automation company. It designs and manufactures automated production, robotics, and process machinery. The firm provides products and solutions to the industries such as meat processing; industrial automation and robotics; appliances; mining; and superconductors. Its business segments are Australasia manufacturing; Americas manufacturing; Europe manufacturing; and China manufacturing. Maximum revenue is generated from the Australasia manufacturing segment. The group operates in New Zealand, North America, Australia, South America, Asia, Russia and former states, Africa and the Middle East, and Other Europe.

Year Founded 1913
HQ Location Dunedin, New Zealand
Last Financing Valuation 43.45
Last Financing Deal Type Merger/Acquisition
Last Known Valuation 43.45
Last Known Valuation Date 30-Apr-16
Last Financing Date 30-Apr-16
Last Financing Size 12.12

Competitors
Harris Waste Management Group

Active Investors
New Zealand Government

Soft Robotics
Website www.softroboticsinc.com

Developer of robotic end-of-arm tools designed to bring automation to high SKU processing, packaging, and supply chain applications. The company’s services include designing and building soft robotic automation systems that can grasp and manipulate items of varying size, shape, and weight, enabling businesses, corporate clients, and industrial corporations to access innovative robotics technology that empowers them to increase business productivity and optimize business efficiency.

Year Founded 2012
HQ Location Bedford, MA
Total Raised 64.32
Last Financing Valuation 125
Last Financing Deal Type Later Stage VC
Last Financing Deal Type 2 Series B1
Last Known Valuation 125
Last Known Valuation Date 29-Jun-21
Last Financing Date 29-Jun-21
Last Financing Size 10

Competitors
Rethink Robotics, IAM Robotics, Magazino, Clearpath Robotics, Savioke, Empire Robotics, MetraLabs, Kindred, RE2 Robotics, Exotec, Grabit, inVia Robotics, Locus Robotics, Bossa Nova Robotics, KNAPP, Fetch Robotics, Applied Robotics (New York), ASIMOV Robotics, 6 River Systems, Perception Robotics, Geek+, RightHand Robotics, iRobot, Robotiq, Swisslog Holding, Mecademic, fellowAI, Optofidelity, Simbe, Fabric (Business/Productivity), SynTouch, Plus One Robotics, Humatics, AutoStore, Skycatch, Osaro, Covariant, Embodied Labs

Active Investors
ABB Technology Ventures, Calibrate Ventures, Fanuc, Four Palms Ventures, Haiyin Venture Partners, Honeywell, Honeywell Ventures, Hyperplane Venture Capital, In-Q-Tel, Mark IV Capital, Material Impact Fund, National Science Foundation, Scale Venture Partners, Taylor Ventures, Tekfen Ventures, Tyson Ventures, United States Department of Defense, Yamaha Motor, Yamaha Motor Ventures & Laboratory Silicon Valley

Frontmatec
Solution in action
The Frontmatec Fully Automatic Cutting Line for Middles is based on optical, servo-electric and mechanical principles.
The machine receives the middles from a primal cutting line or a conveyor system with the neck end first. In the inlet unit, the rib top is pushed against the gripper chain, to ensure the spine is straight and fixed on all joints. Each individual middle is measured with various electro-mechanical devices. Loin and belly are separated with a circular knife. An undercutting knife can be added to separate meat from ribs.

Ensight Solutions
Ensight Solutions
Requires a full line revamp. Lots of downtime and large unit.
Unit needs to be bolted to plant floor, not mobile.

16. Research Flags

Industry Players

Beef Packers
Tyson Foods | Cargill Meat Solutions (Excel) | JBS USA (Swift) | National Beef

Pork Packers
Smithfield Foods | Seaboard Corporation | Tyson Foods | JBS USA | Cargill Meat Solutions

Broiler Chickens
Pilgrim’s Pride | Tyson Foods | Perdue Farms | Sanderson Farms

Outside the United States

Teys Australia
[JBS S.A.](https://en.wikipedia.org/wiki/JBS_S.A.) (Brazil
[BRF S.A.](https://en.wikipedia.org/wiki/BRF_S.A.) (Brazil)
Charoen Pokphand Group (Thailand)
Imperial Cold Storage and Supply Company (South Africa)
Maple Leaf Foods (Canada)
Schneider Foods (Canada)
Cargill Proteins (Canada)
AFFCO Holdings (New Zealand)

Industry Stats

In 2017, the meat and poultry industry processed [7]

9 billion chickens
32.2 million cattle and calves
241.7 million turkeys
2.2 million sheep and lambs
121 million hogs

In 2017, American meat companies produced [7]

26.3 billion pounds of beef
25.6 billion pounds of pork
5.9 billion pounds of turkey
80.2 million pounds of veal
150.2 million pounds lamb and mutton
42.2 billion pounds of chicken

Current Technologies

Chicken Wing and Poultry Automation with mGripAI (Soft Robotics)
60–100 picks per minute, 4 grips per second. Air/vacuum operated grippers with special material patterning to deform and find the right grip size and pressure. No closed loop feedback == increased speed / less complexity. Manually configured grip pressures, 8 configs per profile.
The Soft Robotics Inc. Gripper for Collaborative Robots with Joshua Lessing
500 ham / hr, 20ppl manually, down to 10 ppl with automation. Marketing in EU/JP.
Automatic Ham Boning Robot — Mayekawa HAMDAS-R : DigInfo
From the mfgr: https://www.youtube.com/watch?v=AV2vnFuy8CY
Automatic Chicken Breast Deboning Robot #DigInfo
1500 chickens per hour, 10x the speed of manual ops (? baseline?)
45M JPY / 320k USD