Can AI Help Tyson Foods Lead The Protein Revolution?

Artificial intelligence is forcing car manufacturers to change. Will the same thing happen to Tyson Foods?

Tyson Foods is one of the world’s largest meat producer. Yet Tyson increasingly describes itself as a protein company, rather than a meat company.[i] The Good Food Institute claims that while plant-based meat represents only 1% of the overall protein market, it grew by 23% last year and will continue to increase rapidly.[ii] This is becoming an increasingly mainstream view, as the protein alternative industry has attracted investors such as Bill Gates, Sir Richard Branson, and Eric Schmidt. To seize a portion of the plant-based upside, Tyson opened a venture capital fund in 2016 that quickly acquired a five percent stake in Beyond Meat, the producer of the Beyond Burger, a plant-based burger sold in thousands of restaurants and grocery stores across the world.[iii] Noel White, Tyson’s new CEO, recently acknowledged to the Wall Street Journal that “the marketplace is changing, and we have to change with it.”[iv]

As the firm strives to redefine itself from simply being about “meat” to being a protein company more broadly it may want to consider investing in internal artificial intelligence capabilities, as early signs suggest that algorithms will disrupt the future of protein production. NotCo, a Chilean startup, uses a proprietary machine learning algorithm to create new plant-based recipes that mimic animal-based products. NotCo “mapped 7,000 plants… for their amino acetic structures…that looked like animal-based proteins”[v] and runs that information through its algorithm to predict viable vegan replacements for animal-based products. This reduces the product development cycle by minimizing wasted efforts. Not Mayo®, perhaps the company’s most famous product, is now available in every Chilean Walmart.[vi] Having thousands of plants mapped means that there are effectively an infinite number of possible recipes available to the startup. Without the technology to prioritize R&D efforts, Tyson may find itself behind the protein innovation curve.

In addition to its venture capital arm, Tyson is mitigating the risk of disruption by making significant investments in technology, and has publicly acknowledged plans to develop computer vision systems and other technologies that will improve efficiency (e.g., lower water usage).[vii] Other moves, such as a recent investment in Jerusalem-based Future Meat Technologies,[viii] a clean meat company trying to produce slaughter-free meat at scale, highlight that Tyson is actively seeking new ways to grow and protect its business. Yet there is little available publicly to suggest that the company is using AI to redefine how protein will be produced.

To a degree, this is understandable. While Tyson uses various technologies to improve its core business, it may prefer to leave the dramatic innovations to outside companies. If Tyson felt truly threatened by small companies like NotCo, it could use its venture capital arm to invest in the market and thus gain a portion of the upside. Its investments in Beyond Meat, Future Meat Technologies, and other similar firms suggests this is the company’s strategy. But this presents tremendous risks. In Detroit, after years of ignoring a trend towards autonomous vehicles, carmakers now publicly worry that the bulk of future mobility profits will flow to the makers and owners of software (e.g., Alphabet’s Waymo) while the car manufacturers merely produce the hardware. [ix] Akio Toyoda, CEO of Toyota, told the Wall Street Journal that he seems this threat as “a matter of surviving or dying.”[x] Could this happen in protein production, where the bulk of profits flow to recipe designers, not ingredient producers?

Starting to build an in-house algorithm – or at least purchasing one outright – could reposition Tyson to lead the protein revolution as opposed to relying on its venture capital portfolio and others’ innovations.

I have two main questions about my arguments above:

  • Autonomous cars will rely on data from moving vehicles – something that traditional car manufacturers have a lot of. Yet most of Tyson’s data is likely just about growing animals. Does this data asymmetry mean the comparison doesn’t work?
  • No mainstream research envisions animal agriculture completely ending. Should Tyson simply stick to its core business?

Word count (790 words)

[i] Tyson Foods, “Why We Are Investing in Alternative Proteins.” https://www.tysonfoods.com/the-feed-blog/why-we-are-investing-alternative-proteins, accessed November 2018.

[ii] The Good Food Institute, “The Plant-Based Alternatives Market is Skyrocketing.” https://www.gfi.org/images/uploads/2018/09/Good-Food-Institute-Plant-Based-Nielsen-Data-Sheet-2018-0911-v3.pdf, accessed November 2018.

[iii] Tyson Foods, “Tyson Ventures.” https://www.tysonfoods.com/innovation/food-innovation/tyson-ventures, accessed November 2018.

[iv] Wall Street Journal, “Tyson Foods CEO Looks to Make International Acquisitions.” https://www.wsj.com/articles/tyson-foods-ceo-looks-to-make-international-acquisitions-1541623508, accessed November 2018

[v] TechCrunch, “The Not Company Is Looking To Start A Food Revolution From Chile.” https://techcrunch.com/2018/07/28/the-not-company-is-looking-to-start-a-food-revolution-from-chile/, accessed November 2018

[vi] LiveKindly, “Walmart Chile Now Sells NotCo’s Vegan Mayo Created By Artificial Intelligence.” https://www.livekindly.co/walmart-chile-notcos-vegan-mayo-artificial-intelligence/?mc_cid=3f384c1745&mc_eid=e62f613fdb, accessed November 2018.

[vii] CIO Dive, “Tyson Foods’ Emerging Tech Team Creates A Hatchery For Innovation.” https://www.ciodive.com/news/tyson-foods-emerging-tech-team-creates-a-hatchery-for-innovation/541667/

[viii] Bloomberg, “Tyson Foods Makes Another Investment in Lab-Grown Meat.” https://www.bloomberg.com/news/articles/2018-05-02/u-s-food-giant-tyson-makes-first-investment-in-israel, accessed November 2018.

[ix] Wall Street Journal, “Will Tech Leave Detroit in the Dust?” https://www.wsj.com/articles/can-detroit-become-a-software-business-1540008107, accessed November 2018.

[x] ibid

 

Previous:

Sorting and securing at scale: machine learning at Dropbox

Next:

Machine Learning at Airbnb

2 thoughts on “Can AI Help Tyson Foods Lead The Protein Revolution?

  1. While one could justly question Tyson for being a little behind the AI 8-ball, NotCo’s approach here seems very interesting. By relying on existing data sets for plant proteins, they can use the algorithm to essentially run simulations of recipes and test for taste, texture, nutritional content, etc. This could be a very inexpensive way to probe for new protein alternatives. I would think a conglomerate like Tyson could either purchase recipes from companies like NotCo (similar to a pharma model) and become a mass producer or try to take this development in house.

  2. Great article! For me, where the analogy breaks down is that self-driving cars are still produced like cars but plant-based and clean meat are largely not produced like traditional animal protein. For this reason, I think today’s traditional auto OEM’s would provide much more value in a future self-driving world than today’s Tyson would provide in a future alternative protein world. Sticking to a core business that is declining isn’t a viable option, in my mind. Therefore, I believe that Tyson has no choice but to start learning how to develop and produce plant-based and clean meat.

Leave a comment