Talking AI With Samuel Bertram, CEO & Co-Founder, OnePointOne
What’s in a name? For brothers Samuel and John Bertram, it signifies nothing short of a desire to improve the human condition by revolutionizing agriculture through automated indoor farming. Their company, OnePointOne, is on a mission to nourish and heal humanity by unleashing the power of plants.
Born in Melbourne, Australia, the duo came to the US in the 2010s on collegiate tennis scholarships. Over the years, they honed skills off the court– including electrical, mechanical and robotic engineering — and began searching for business opportunities that could affect large-scale, positive changes on human health, the environment and agriculture.
Galvanized by the fact that 1.1 billion people began this millennium malnourished, Samuel and John co-founded OnePointOne, Inc. in 2017. Serving as a constant reminder of what they’re aiming to solve, aptly named OnePointOne has developed an automated, aeroponic, indoor farming system to grow fresh food in urban areas around the world. OnePointOne’s proprietary technology now powers Willo, the company’s new consumer brand, which launched earlier this year. Willo’s Farming as a Service (FaaS) subscription model is designed to reconnect people and families directly to the farm and the initial response has exceeded all expectations.
In this month’s Indoor Ag-Con Q & A, OnePointOne Co-Founder & CEO Samuel Bertram shares his thoughts on opportunities for the vertical farming industry, emerging AI innovations and plans for his own growing business.
When one thinks of AI for indoor agriculture, what are some of the key areas of need that indoor growers have today?
AI refers to the developing ability for machines to replicate human decision-making and behavior. That said, areas for useful AI development include:
- Plant Health – Using AI to determine the health status of any plant by comparing large—predominantly imagery—data sets against in-situ imagery: disease detection, photosynthetic health, etc.
- System Monitoring – Beyond direct sensor readings, large data sets of factors like CO2, fluid flow, fluid pressure, temperature, and others, can be used to determine more anomalous malfunctions of the system
- System Optimization – Large data sets describing the life experience of the plant, i.e. light levels, temperatures, and humidity, can be used to improve the performance of the system to any cultivar.
By taking in large quantities of data from a variety of environmental, system and plant sensors, AI techniques can be applied to optimize performance of the vertical farming system and assist growers in their role as farm operator. While traditional sensor readings like temperature, humidity and flow rate are vital, plant imagery is a requirement to unlock AI’s power in vertical farming. Without high-frequency, high-resolution, hyper-spectral imagery of all plants in vertical farm production, the vertical farming industry will never reach its full potential. OnePointOne has focused heavily on collecting, analyzing and providing that imagery data to our growers.
It’s important to note that while AI can handle far more data than a human being, humans are still superior when it comes to complex decision-making in most cases. Therefore, AI should be focused on deriving learnings from massive data sets, informing growers of those learnings, and unlocking the potential of the system and the grower.
Cost is a critical component to any indoor farm operating, hopefully, profitably. Please share some idea of the cost spectrum (low to high) when one considers the implementation of ANY AI technology system within an indoor farm.
I would break this down as follows:
1. Labor is the highest operating expense cost inside of a vertical farm. Therefore, vertical farms must optimize their utilization of labor.
2. After automating processes like seeding, plant movement, and plant harvest/packaging, vertical farms must solve the problem of system and plant monitoring costs.
3. If system and plant monitoring become automated, high-skilled labor can be centralized.
4. Centralizing high-skilled labor dramatically improves the cost and scalability metrics of vertical farming.
5. Without high-frequency, high-resolution, hyper-spectral imagery of all plants within a production system, centralization of high-skilled labor cannot occur.
6. Without this significant improvement in labor utilization, vertical farms will continue to struggle for profitability.
The highest-impact application of AI in vertical farming is through the analysis of environmental, system and plant-imagery data-points and their corresponding impact on plant yield and quality.
Then, AI can “automatically” improve the quality and yield of biomass, while optimizing the use of resources, i.e., light, HVAC and irrigation.
Optimizing the usage of electrons for lighting, temperature and humidity control is the perfect job for AI. This will dramatically improve the economics of vertical farming over time.
You are a Bronco from the University of Santa Clara. How can higher education, top horticulture universities and R&D institutions help accelerate AI technology in indoor vertical farming?
I see three clear ways these institutions can help:
1. Cultivating high-quality minds that will push our industry forward
2. Conducting vast numbers of experiments to develop AI algorithms to detect optimal and suboptimal plant health (potentially high-throughput phenotyping, for example)
3. Licensing state-of-the-art vertical farming technology to standardize production in pursuit of standardizing data
Given OnePointOne’s location in the heart of the Silicon Valley, are you seeing any large tech firms getting involved with AI for indoor horticulture? If so, who?
Absolutely. Two come immediately to mind:
Google is diving deep into imagery-data analysis for outdoor farms. It is only a matter of time before they venture into the vertical farming space.
Amazon (AWS) has developed several teams and tools that can assist vertical farms in their storage and analysis of data.
What are some of the AI advancements OnePointOne is focused on now?
Imagery, imagery, imagery. Similar to Tesla, high-frequency, high-resolution, hyper-spectral imagery data sets from production and research farms will give OnePointOne a sustainable competitive advantage. Beyond real-time image analysis — like leaf-edge detection, discoloration detection and discontinuity detection, etc — large imagery data sets will drive system optimization unlike anything else. With enough imagery data, software models of plants can be created, then used in production settings to ensure optimal plant performance.
Above all else, OnePointOne is focused on plant-quality. Employing AI to our imagery data allows us to constantly improve the quality of our plants, while minimizing the required input resources.
Last, but certainly not least, let’s talk about this year’s launch of your consumer brand Willo. What was the rationale behind this subscription model and what type of response has it received?
For ten thousand years farming was local and provided a diverse range of nutrients for the community. Today, we have no idea where our food is coming from and we are losing the nourishment battle. Willo exists to reconnect us to the farm and ensure that we are nourished by the highest quality foods imaginable.
Willo’s mission is enabling personalized plant-based nutrition to optimize human and environmental health. We achieve this through Willo’s personalized farming service, which allows members to control their own farm plot and regain access to locally grown, high quality produce unlike ever before.
Within a matter of weeks of our launch, Willo sold out the first farm. We are now in the process of building a farm ten times larger to accommodate the building waitlist for our personalized farming service.