In addition to labor costs, labor shortages have also been a long-standing trend in California agriculture as American agriculture increasingly competes with Mexican farms for farm labor . This structural shift in the labor pool has caused fewer workers to be available for low-skill, low-wage agricultural jobs. Furthermore, human labor also is time-consuming and error-prone, eliminating only an average of 65 % to 85 % of weeds . To address the problems associated with conventional weeding methods, such as greater herbicide resistance and labor shortages and costs, the emergence of precision weeding technologies has become more prominent, supported by recent advancements in sensing technologies and artificial intelligence. Precision weeding, targeted and selective removal of weeds using advanced technologies and equipment, has emerged as a compelling solution to address the limitations of conventional weeding methods in commercial agriculture. This method encompasses various techniques, including mechanical, chemical, and thermal/electrical weeding, presenting a versatile toolkit for growers. Many promising prototypes of precision weeding, covering all aforementioned techniques, have been achieved in academic settings . In addition, vertical racks early commercial operations have already proved the value proposition of fulfilling labor needs. One mechanical cultivator reduced the weed density by 27 to 41 % more compared to the conventional option .
The success of the precision weeding sector is dependent on the compatibility between the startups’ value propositions and the growers’ on-the-ground needs, which in this study is examined within the lenses of objectives, stakeholder interactions, and grower user journeys. This compatibility is heavily impacted by social systems and institutional contexts and individual grower perceptions and characteristics. In a USDA study about the adoption of precision agriculture for commodity row crops like soybeans, researchers found that drivers of adoption include grower participation in USDA programs . Precision weeding adoption has been encouraged by the USDA Conservation Stewardship Program, which provides contract payments for environmental improvements. Technology adoption systems are often highly localized, with stakeholders optimizing their interactions with one another through curated local ecosystems. A study focused on institutional factors that impact startup formation uncovered significant relationships between most county-level variables while much less significant ones between states, implying the importance of local ecosystems . Beyond startup formation, localized networks for innovation can facilitate participatory, co-learning relationships between technology providers and technology users. Case studies of the adoption of decision support systems have shown that these technologies serve as a neutral meeting ground between scientists and growers, enabling the social context of co-learning cycles to influence individual learning .
Government-funded initiatives, as well as local ecosystem dynamics, help shape grower willingness. Individual grower characteristics will also influence the success of precision weeding commercialization. Larger farm sizes are correlated with reduced risk aversion, the specialization of managerial labor, and lower per unit costs for equipment . These characteristics lend to larger farms having a higher level of willingness to adopt new, riskier precision agriculture technologies. In addition, farmers also highly valued financial considerations such as economic returns and capital investment sizes. Characteristics that may impact whether Californian farmers adopt new technologies are their ages and social networks . Though adoption rates are also positively correlated with older age, more experience, and the affordability to take risks, younger farmers may experience the lower cognitive cost of switching and thus be more likely to adopt new technologies . Furthermore, social networks impact the probability of adoption. This correlation can be exhibited through an inverted U-shaped in which the probability of adoption is low when only a few people in the network have adopted or almost all have adopted, but high when about half of the people in the network have adopted . Thus, word-of-mouth could be the key factor in growers becoming aware of precision weeding technology, considering its adoption, and making an adoption decision. Precision agriculture technologies are often used in tandem.
Therefore, precision weeding technology adoption rates could be higher for farmers who also use technologies such as GPSmapping and guidance systems . Grower characteristics and decision-making processes cannot be fully understood without context on California growers and the other stakeholder groups involved in technology commercialization: startups and venture capital firms. In 2019, California had 69,900 farms with an average farm size of 348 acres . This average farm size is smaller than the national average of 445 acres . In 2017, Californian farmers had an average age of 59.2 years, and less than 6 % of all California producers, defined as farmers and ranchers, were aged 24 or younger . Most farms–74 %–were owned by individuals, families, or partnerships. Although 10 % of farms were owned by corporations, less than 2 % were owned by non-family corporations . Most farms are operated by California’s 700,000 seasonal farmworkers . Our study design is limited to growers operating in Central California because of its dominance in Californian and American agriculture. The Central Valley produces over 250 crops with a value of $17 billion per year, an estimated 25 % of the nation’s food, and in 2017, the Central Coast had over 400 vegetable farms and nearly 800 fruit and nut farms . Startup companies are another key stakeholder playing a significant role in California’s precision weeding ecosystem. Precision weeding startups have increased in popularity in the past few decades and have a positive growth trajectory, with the sector expected to grow by $268.75 million with a CAGR of 18.41 % from 2022 to 2026 . Precision weeding startups may be complementary to existing R&D efforts at larger corporations or seeking to disrupt current technologies or business models, potentially eroding monopoly power . However, incumbent agribusiness has retained their advantages of controlling supply chain and distribution standards. In addition, many startups seek out partnerships with incumbents for legitimacy and practical assets, such as manufacturing capacity . While startups often emphasize their value-based approaches and product innovation, they are inhibited by limited resources and look towards incumbents for process innovation . Despite some startups toning down their ambitions after working with incumbents, many argue that VC-backed startups are a more efficient use of capital in comparison to global corporate agriculture R&D expenditures . Investments from venture capital firms enable startups to take on the risky endeavors of developing and trialing novel technologies. In exchange for taking the chance on and providing capital to startups they ascertain as having exponential growth potential, VCs receive equity in the startups. VC investments in agri-food startups have accelerated in the past decade, with “twenty times more capital…invested in new agtech ventures in 2021 than 2012 .” The average seed round for digital and precision agriculture startups is $3.8 million and the average funding increases to $117 million for Series D and later funding stages . Up to 2006, global investments in agtech startups remained less than $200 million per year but then steadily grew until investments began exceeding $3 billion every year in 2010 . The large VC investments in agtech are a reflection of their general ‘pioneering,’ ‘adventurous,’ and ‘disruptive’ culture .
California is a natural geographical focus for our study considering the state’s agtech innovation ecosystem has strong interparty dependence, cannabis drying curing the availability of venture capital, and a forward-thinking and tech-savvy consumer market . In addition to the three stakeholder groups directly involved in commercializing new technologies, government agencies are incentivized to be involved in the agricultural industry for economic reasons. In 2021, California’s top twenty commodities produced over $44 billion in value . For crop commodities, California regulates pesticide use through the California Department of Food and Agriculture and the California Department of Pesticide Regulation. The government historically played a more significant role in agricultural research and development through supporting agricultural research stations and academia as well as supporting philanthropic foundations . Outside of direct research and development financing, the public sector also serves as allies in cleantech startup innovation, particularly with licensing alliances with universities and developing markets through demand-pull policies . In addition, governments play a large role in the mobilization phase in particular because they select participants and set the criteria for funding . With regards to precision weeding, the Governor’s Office of Business and Economic Development has tax credit and sales tax exemption programs for the agriculture and agtech sectors. In the U.S., the USDA, other federal agencies, and state governments all administer public agricultural R&D funding . Although existing research has explored how growers adopt technological advancements through collaboration with commercialization partners, there has been limited focus specifically on the precision weeding sector. The growers’ user journey could be different depending on the startup’s business model . Central California specialty crop growers face unique challenges compared to row commodity growers because specialty crops only account for 10 % of the U.S.’ farm operations but shoulder the complexities of higher risk, variable costs . A crucial gap exists within the body of literature about adoption networks regarding the relationships between stakeholders in the precision weeding ecosystem. Addressing this gap is imperative for elevating the advantages of precision weeding to the forefront of modern agricultural practices and comprehending the commercialization of precision weeding technologies, defined as the entry of such technologies into the mass market. This research study was motivated by the need to better understand how the relationships between stakeholders in the precision weeding ecosystem impact the adoption of the technology. While other stakeholders, such as intermediary organizations like cooperatives, play a role in technology proliferation, we limited the scope of our study to the core groups directly involved in commercialization. The objectives were defined to address specific aspects of these relationships: compatible motivations, investigating whether there are shared motivations among stakeholders, collaborative models, examining the effectiveness of existing collaborative models between stakeholders, and user journey, analyzing the user journey for growers adopting precision weeding technology and how technology adoption is impacted by other ecosystem players.The data collection method employed for gathering textual data to answer the three objectives involved semi-structured qualitative interviews. This method was chosen over other qualitative methods, suchas surveys and focus groups, because open-ended exploratory questions best suit the perception-based and subjective research questions. Though the number of qualitative interviews required to draw legitimate conclusions is contentious, it is widely accepted that most novel information is generated early on in the data collection process in an asymptotic curve and then there is a steep drop off in new information. Analyzing three datasets—the first with 40 interviews and 93 unique codes, the second with 48 interviews and 85 codes, and the third with 60 interviews and 55 codes—researchers found that six to sixteen interviews could reach a median degree of saturation of 69 to 89 %. . Therefore, we conducted 17 interviews to comfortably predict high data saturation. Saturation assesses the rigor of qualitative sample sizes, indicating when additional data adds little or no new information to answering the research questions. Of the 17 interviews, we conducted interviews with seven Californian growers, five venture capital firms/accelerators, four precision weeding startups, and one government agency. A small sample size was effective in revealing the core categories of these lived experiences . Regarding participant recruitment, we used purposive sampling to certify that interviewees had experience within the precision weeding sector and held mid to high-level roles in their respective organizations, ensuring the validity of the data collected. The outreach process was through networking at a relevant in-person conference, cold emailing and social media messaging, and speaking with ‘connectors’ such as UC Cooperative Extension Specialists. In addition, we used the snowball sampling technique, whereby we asked interviewees to recommend additional study participants. Due to logistical constraints, the interviews were conducted over video calls. Prior to the interview and following an explanation of the study objectives, informed consent was obtained via a form through WeSignature or verbally, allowing for voice recordings and transcriptions of the interviews. A caveat is that two of the interviewees’ informed consent forms did not allow voice recordings due to their companies’ legal directives but did allow for the interviews’ content and quotes to be used in this research study. The voice recordings were processed through the software Fireflies.ai to acquire transcripts, which were manually proofread following the interviews. The number of questions asked was based on data and thematic saturation, a criterion used for discontinuing data collection and analysis .