Farm-to-Table 2.0
How Farmers + Restaurants Can Unlock Value in a Changing World and Value Chain
While still early innings, the food-tech space finds itself poised for a seismic transformation. The advent of AI/ML applications creates opportunity for farmers and restaurants to better utilize finite resources and capacity in a world where demand for costless and convenient consumption is the new norm.
According to a UN report, it’s estimated that the world’s population will reach 9.8 billion in 2050. To meet this growing base, “…global food production will need to increase 25–70% by 2050.” (Refresh “Soil to Supper” Report)
Further exacerbated by climate change, “We have to grow more food in the next 30 years than we have in the last 8,000 with less land and fewer people desiring to become farmers.” (Refresh “Soil to Supper” Report)
From the consumer side of things, according to the WSJ and PitchBook, “Venture-backed delivery platforms including DoorDash Inc. and Postmates Inc. raised about $14 billion across 198 deals from 2016 through 2018.” That doesn’t include DoorDash’s recent ~$400MM raise.
So, why does this all matter?
“Off-Premise Operations” & Outsourcing in Restaurants
To more cost-effectively meet rising online demand, restaurants are leveraging third-party ghost (“cloud”) kitchens and food delivery services, outsourcing much of their operations (e.g., preparation and delivery).
While this provides shorter-term cost advantages associated with the necessity to capture an uptick in demand, I envision traditional restaurants that operate this way to resemble more of a luxury than a need in the future. Fast-casual joints like Zume Pizza and Spyce, a dining class of increasing popularity, have elected to build a new model from scratch, using automation and robotics to lower labor/production costs. Sweetgreen has done something similar with an immensely successful online platform.
Even still, some ghost kitchens (think: Keatz in Europe) have already utilized learnings to forward integrate, offering custom meals direct-to-consumer, thus bypassing the restaurant completely. While this level of personalization is likely to persist as we trend towards an “at-home” food economy, bolstered by online restaurant delivery sales expected to grow to $62B by 2022, I don’t see vertical integration as the answer (case in point: failure at Munchery & Sprig). Rather, I see platform-based approaches that take large swathes of data and apply it to the customer experience, feeding the flywheel.
Take Tastewise (raised a $1.5M Seed in May 18'), for example. With a platform able to analyze billions of food data points, from menus and recipes to social media activity (i.e., social listening), the company is able to provide food producers with key insights related to customer preferences and emerging trends. Companies like Spoonshot have a similar value prop — so I’m trying to better understand where Tastewise is pulling data from/what makes it defensible. A further extension is Shelf Engine’s (raised $5.2M in total funding) prediction model that allows grocers to meet daily demand without having to worry about the cost of wasted food.
Leading Indicators, Collaboration, and Smart Devices in Farming
For much of its history, farming has depended on variables largely out of its control. Weather patterns were unreliable and the onset of diseased crops were more of a lagging indicator, than leading, costing farmers significant revenue during harvest. This is now changing. AI’s dissemination across different data applications has provided farmers with an ability to spot symptoms earlier and detect disease in crops/livestock before incurring irrevocable damage to surrounding crops, improving yields and underlying profitability. The advent of big data paves the way for farmers to integrate smart systems (via autonomous technology, drones, and sensors) across all aspects of the farm, using insights gathered to compare/collaborate with other farmers, improving the overall efficacy of the system.
Companies like Cainthus have partnered with Cargill to bring computer computer vision and AI to translational imaging within livestock. HerdDogg (raised a $3M Seed in July 18') seems to be up to something interesting here with its livestock monitoring platform that continuously tracks livestock health, producing roughly 100M rows of data/year for small farmers. HerdDogg adds a collaborative element to the platform, allowing farmers to compare their own livestocks’ health with that of other farmers in the area. However, I’d like to dig deeper on the speed with which meaningful results are provided to farmers, as well as the efficacy of the results. Would computer vision technology, as Cainthus is using, prove to be more or less effective?
Farmwave, a cloud-based platform that uses AI to detect diseases across crops within seconds via camera phone (with ~90% accuracy), seems to be on to something exciting. Using IoT connectivity, the app gives farmers instant access to data pertaining to their farmland and the ability to connect with other farmers in the same locality or abroad, potentially preventing disease outbreak at other farms. I’m wondering how Farmwave’s accuracy compares to that of the industry and image recognition technology at large.
All-in-all, I am deeply excited for what the future of food-tech holds. I find that we face a world on track to “demand more, with less” and am eager to work with entrepreneurs hell-bent on solving this problem.
Onward and upward,
Jackson Feder
