Fourth Agricultural Revolution: Who bears the costs, who gains the benefits?

Today’s science policy conversation is mostly about the urgent need to facilitate a Fourth Agricultural Revolution driven by new digital and genomic technologies (DNAI). The emphasis is on strengthening and diversifying food supply chains, bringing “cutting edge” digital information and communications directly to food producers, and enhancing resilience and diversity in crop and livestock production. This Fourth Revolution is generally presented as accessible and beneficial to producers everywhere regardless of size or strategy and that additional tools are equally compatible with agroecological or industrial production models. However, as with any major innovation – especially those affecting food security – it is crucial to examine who ultimately bears the costs, who reaps the benefits, and whether this pathway fosters inclusivity or reinforces inequality. In this post, I will explore these questions, using examples from both the agricultural and technological spheres to shed light on the complex trade-offs involved.

Food in a digital and artificial intelligence (AI)-driven ecosystem

In annexing itself to this fourth Revolution, food – more than ever – becomes an economy class passenger on a multi-class Airbus that includes much larger sectors of the global economy dominated by an oligopoly of trillion-dollar enterprises who see it as nothing more than another not especially profitable widget competing for their band with, storage space and satellite attention. This has real implications. Just as bulk cargo vessels were pushed aside by container ships going through the Panama Canal during last year’s drought, services for the diverse and diffused food system will be subordinated to more valuable customers.

The fragility of digital dependency

Digital technologies are far from invincible: satellites crash; blockchains break; data is jettisoned crypto currencies evaporate; and long supply chains unravel. Over time, some of these problems might be overcome – but each might also be replaced by another layer of unanticipated techno-failures. For food producers and the food insecure, “overtime” always comes too late. Digital technologies therefore have potentially high hidden costs.

Lessons from other sectors on the issue of supply chain vulnerabilities

It is instructive to examine the impact of new technologies on the supply chains of other, more powerful, economic sectors. For example, it took war in Ukraine for NATO and the arms industry (the most risk-averse and astute industrial sector) to realize that “just in time” and “just for practice” munitions supplies were inadequate for prolonged ground warfare. Secondly, during Covid, vaccine manufacturers belatedly realized that mRNA vaccines needed more than 2000 inputs including difficult to obtain proteins from a moth in India*1 and an adjuvant only seasonally available from a Peruvian soapbark tree.

Thirdly, although most governments are aware that one US company, Nvidea, is the sole designer and patent holder of the H100 top-grade AI chips, it was only when serial hurricanes struck the US Southeast late in 2024, that Silicon Valley, discovered that at least 70% of the world’s ultra-fine quartz (i.e. sand) required for the crucibles that make nanochips comes from a private Belgian firm, Sibelco, whose sole mine (now indefinitely out of order) is in mountainous North Carolina which has, in the last few years, also been shut down by a fire and another flood*2. Investors only then became aware that the highest performance AI chips are exclusively manufactured by TSMC in Taiwan that has been thwarted by local droughts and Covid labour shortages in 2022 – 23 that, in turn, can only buy its highly-specialized fabrication machinery from ASML, a Dutch company whose sole plant outside of Berlin suffered a serious fire in 2022 that again, in turn, depends on the supply of unique lasers from MKS, a privately-held US company, whose data was held ransom by cyberhackers in 2023.

The true cost of these vulnerabilities is opaque to the Global South’s food producers and governments who must get their information from Elon Musk’s problematic Starlink satellite arrays. The Global North, meanwhile, uses transoceanic cables for 94% of their data largely unaware that, again, only a handful of giant companies own and are capable of repairing the cables and so few companies – dominated by Amazon, Microsoft and Alphabet (Google) and Meta (Facebook) – control the world’s data centres.

The Fourth Agricultural Revolution: a challenge for local food systems

The greatest concern about the innovative style of the Fourth Agricultural Revolution is that it is in no way an entry to more local and national food systems since its cost and size require it to dominate the field and subordinate other forms of innovation. Although it can be argued that access to global information and global supply chains can be a “backup” to local food systems, the practical outcome will be otherwise. Food access and food production are too important to our societies to be out of reach. The Third Agricultural Revolution is already rendering it practically or contractually difficult for producers to bargain with riparian behemoths to replant seed or repair farm machinery. Removing production data and reducing the capacity to innovate in a Fourth Revolution simply exacerbates the dangers of food insecurity in a century of climate and biodiversity crisis. For example, an estimated 45% of all private sector agricultural research is focused on one crop: maize. Almost all private research and development (R&D) is devoted to only five crops and five livestock species while small-scale producers breed – and plant – almost 7000 crop species and more than 30 livestock species.

Localised research and development are a source of strength

Recent studies indicate that R&D labs operating adjacent to production facilities produce more patents than those distant from production sites. Secondly, there is evidence that there are significantly more citations of peer-reviewed papers authored by proximate collaborating scientists than by top-ranking scientists collaborating at a distance. Thirdly, the R&D divisions of independent enterprises linked in multi-tiered supply chains invariably seek innovations that improve their own profitability with little or no reason to strengthen resiliency along the entire chain. Finally, economies of scale encourage R&D units to prioritize high-tech innovations*3 that have global applications over wide-tech applications*4 that prioritize community/ecosystem-appropriate strategies.

The need for sustainable research and development

Contrary to popular assumptions, US private sector innovation shows inconsistency due to mergers, acquisitions, and erratic Head Office priorities. Start-up companies, especially, are experiencing a steep decline: while there were 344 new unicorn startups in 2021, there were only 45 in 2023. Investment in start-ups was cut by half from USD350 billion of new money during 2021 to USD170 billion in 2023*5. Studies also show a 15% drop in start-up patent applications. The sunk costs of abandoned startups and wasted scientific work years is incalculable. To be clear, however, although startups are suffering, the major AI players (Microsoft, Google, Amazon, Meta, OpenAI and Apple) are expected to invest USD200 billion in 2025 mostly building data centres and more powerful Large Language Models (LLM), but many regard this as money spent on market control, not on groundbreaking research.

Responding to crisis: local research and development as a resilient alternative

Climate and biodiversity crisis will play out differently in different social contexts and ecosystems. Large-scale R&D systems have no incentive to – and can’t respond to – local risks or take advantage of local human and material resources. Conversely, local communities or ecosystems can efficiently bring together complementary strategies supporting food, health, housing, labour and childcare/education (especially for most-affected populations). Farmers, suppliers, transporters, retailers and consumers are constantly rubbing shoulders and exchanging information. Secondly, solutions produced by industrial R&D systems discount risks involved in long, complicated or politically pragmatic supply chains, whereas local innovations consider and build upon local supplies and resources. Additionally, local innovation systems see these advantages of piggybacking on trusted wider middle and even long-distance networks.

Moving forward: prioritizing local innovation in the Fourth Agricultural Revolution

The bottom line is that the Fourth Agricultural Revolution is seeking dominance based on the unproven assumptions that “big” is better and that “new” is the exclusive domain of global technologies. In reality, this century’s global threats in climate and biodiversity can best be addressed by local responses – always open – and actively seeking – wider, trustworthy information and partnerships – but anchored in immediate realities.

*1 Novavax is a COVID-19 vaccine that replaces problematic polyethylene glycol proteins with cells from moths (Spodoptera frugiperda).

*2 An intensely private Belgian company, Sibelco, owns a mining operation in Spruce Pine, North Carolina overwhelmingly dominates (70 – 90%) the ultra-fine quartz silica (a.k.a. “sand”) used to make the world’s highest-quality (3 nm–5 nm) computer chips. The mine experienced a major flood in the 1990s.

*3 High-tech refers to advanced, complex technologies that push the limits of innovation, typically aiming for efficiency, scalability, and universal applicability. These solutions often require significant resources and specialized infrastructure, making them suitable for large-scale, global use. Examples include AI, gene editing, advanced robotics, and space technology.

*4 Wide-tech solutions prioritize accessibility, adaptability, and sustainability within local or specific contexts. They are designed to be user-friendly, resource-efficient, and appropriate for community or ecosystem-based needs. Examples include local water purification systems, open-source software for education, and decentralized renewable energy solutions.

*5 Unicorn arrivals and investment fluctuations may or may not indicate a long-term trend. The important point is that good science – and important technologies – develop best in stable research ecosystem and such erratic market behaviour demoralizes scientists and disables innovation.