SWITCHED ON
The daily technology series nobody asked for but everyone needed
Feed the World 2.0: Food Security and the Precision Agriculture Revolution
By 2050 the world needs to produce roughly sixty percent more food for a population that will be larger, wealthier, and demanding more protein, on land that is increasingly stressed by climate change and water scarcity. Technology has answers. Not all of them are the ones being advertised.
The Green Revolution of the 1960s and 70s — new seed varieties, fertilisers, irrigation — averted the mass famines that were confidently predicted by the best demographers of the era. It is the most important underreported achievement in the history of human welfare. The question for the next thirty years is whether a second such revolution can be engineered, on a tighter timeline, against a climate that is working considerably harder against us than the one the first revolution operated in.
Yesterday we looked at spatial computing through the lens of what Apple Vision Pro's reception actually tells us about interface transitions, the enterprise and industrial use cases finding genuine traction, and digital twins as the quieter and more economically significant application of the underlying technology. Today we are back to immediate stakes — as immediate as it gets, actually, because we are talking about food. Specifically, about the convergence of climate stress, population growth, water scarcity, and the agricultural technology revolution that is simultaneously the most important and least discussed technology story of our era. Precision agriculture, drought-resistant crop engineering, AI-driven farm management, and the real constraints that none of the Silicon Valley optimism about feeding ten billion people tends to adequately acknowledge.
01 — The Problem in Numbers
The global food security challenge is not primarily a technology problem. It is a distribution problem, a political problem, a trade problem, and increasingly a climate problem, with technology as one of several necessary but insufficient responses. Getting this framing right matters, because a technology-only narrative produces the wrong policy conclusions and, historically, has left the most vulnerable populations behind while the efficiency gains flow to large commercial producers in wealthy countries.
With that caveat front-loaded: the numbers are genuinely alarming. The global population is projected to reach approximately ten billion by 2050. Dietary transitions in rapidly developing economies — more protein, more calories, more processed food — mean per capita food demand is rising faster than population alone. Climate change is reducing crop yields in regions already under food stress: Sub-Saharan Africa, South Asia, and parts of Latin America are facing combinations of heat stress, drought, and shifting rainfall patterns that are already reducing yields of staple crops in some areas. Global freshwater for irrigation — which accounts for approximately seventy percent of all human freshwater use — is under increasing pressure from overextraction and changing precipitation. The topsoil that took thousands of years to accumulate is being lost through erosion at rates that conventional agriculture cannot sustain indefinitely. The UN Food and Agriculture Organization estimates that roughly a third of all food produced globally is lost or wasted before it is eaten. Any serious accounting of the food security challenge has to grapple with all of these simultaneously.
02 — Precision Agriculture: The Data-Driven Farm
Precision agriculture — using data, sensors, GPS, and increasingly AI to manage farms at much finer spatial and temporal resolution than traditional practice — is the most mature and commercially deployed agricultural technology category, and its impact is already measurable in farms that have adopted it.
The core insight is that traditional farming treats a field as uniform when it is not. Soil composition, moisture, nutrient levels, and pest pressure vary significantly across even a single field, and treating the whole field identically — applying the same rate of fertiliser, the same irrigation volume, the same pesticide application — wastes input and produces suboptimal yields. Precision agriculture installs sensors in the soil and on equipment, uses satellite and drone imagery to map crop health at centimetre-level resolution, and applies variable-rate inputs — more fertiliser here, less there, irrigation only where moisture sensors indicate the need — guided by AI analysis of the sensor data.
The documented results in farms that have adopted precision agriculture consistently show reductions in fertiliser and pesticide use of fifteen to thirty percent, with maintained or improved yields. Reduction in input use is simultaneously an economic benefit to the farmer, an environmental benefit through reduced chemical runoff into waterways, and a contribution to greenhouse gas reduction since nitrogen fertiliser production is a significant emissions source. These are real gains that are already being captured by commercial agriculture in wealthy countries. The technology access gap — the cost and technical sophistication required to implement precision agriculture — means smallholder farmers in low- and middle-income countries, who produce a significant fraction of the world's food, have largely not benefited. Closing that gap is a policy and distribution challenge more than a further technology challenge.
The most immediately impactful agricultural technology investments may not be new seed varieties or drone fleets. They may be soil sensors, connectivity, and the extension services needed to translate sensor data into decisions that smallholder farmers in Sub-Saharan Africa and South Asia can actually act on. Less exciting to investors. More likely to feed people.
03 — Drought-Resistant and Climate-Adapted Crops
Plant breeding has always been about selecting for desirable traits, and modern biotechnology has dramatically accelerated the pace at which new varieties can be developed and tested. Conventional selective breeding operates on generation timescales — years per cycle. Marker-assisted selection, which uses genetic markers to identify plants carrying desired traits without waiting for them to express phenotypically, compresses this. CRISPR gene editing, applied to crop development, can introduce specific targeted changes in weeks rather than the years required for traditional breeding programs. The combination has produced a pipeline of drought-resistant, heat-tolerant, and disease-resistant crop varieties that is substantially larger than at any previous point in agricultural history.
Several are already in the field. Drought-tolerant maize varieties developed through conventional breeding and marker-assisted selection have been deployed across Sub-Saharan Africa through partnerships between CIMMYT, the international maize and wheat research centre, and national agricultural systems, with documented yield improvements of twenty to thirty percent in drought years compared to local varieties. Flood-tolerant rice varieties carrying the Sub1 submergence tolerance gene have been adopted across millions of hectares in South and Southeast Asia, providing critical resilience for smallholder farmers in flood-prone areas.
The CRISPR pipeline for crop improvement is earlier in development but addresses problems that conventional breeding struggles with. Varieties resistant to specific diseases — wheat blast, a devastating fungal pathogen expanding its range as temperatures rise — have been developed using CRISPR. Varieties with improved nitrogen use efficiency, which would reduce fertiliser demand while maintaining yields, are in development at multiple research institutions. The regulatory treatment of CRISPR-edited crops varies enormously by jurisdiction: the US and several other countries treat CRISPR edits that could have occurred through conventional mutation as not subject to GMO regulations, significantly accelerating approval. The EU maintains a more precautionary approach. This regulatory divergence is itself a food security consideration, because it affects which technologies reach which markets and on what timescale.
04 — AI on the Farm
Artificial intelligence is entering agriculture across multiple dimensions simultaneously, from satellite-based yield prediction at national scale to individual plant health diagnosis from smartphone photos. The breadth reflects both the genuine applicability of machine learning to pattern recognition in agricultural data and the somewhat feverish enthusiasm for AI applications in every sector that has characterised the past several years.
The most clearly validated current applications are in computer vision for crop disease and pest identification — apps that allow a farmer to photograph a diseased leaf and receive a diagnosis and treatment recommendation have been deployed across Sub-Saharan Africa and South Asia through platforms including PlantVillage and Plantix, reaching millions of smallholder farmers who previously had limited access to plant pathology expertise. The accuracy of these systems in controlled studies is high for common diseases of major crops, with limitations in novel disease identification and in providing context-sensitive advice for the full range of local conditions and available inputs.
AI-driven irrigation scheduling — using weather forecasts, soil moisture data, and crop water demand models to optimise irrigation timing and volume — has produced documented water savings of twenty to forty percent in trial deployments, a result that is significant given that agricultural water use is the largest single freshwater demand in most countries. Autonomous farm equipment — robots that weed, thin, and harvest with precision that human labour cannot match at equivalent cost — is progressing through commercial deployment for specific high-value crops including strawberries and lettuces, where the economics of automation are most favourable. Extension to staple crops at global scale is a longer horizon.
05 — The Distribution Problem Technology Cannot Solve
The world currently produces enough calories to feed everyone on the planet. Hunger exists not primarily because of insufficient production but because of distribution failures, conflict, poverty, and the political economy of food systems that consistently direct food where purchasing power is concentrated rather than where nutritional need is greatest. Technology that increases global food production without addressing the distribution architecture of the global food system is technology that produces more food for people who already have enough of it.
This is not an argument against agricultural technology investment. It is an argument for holding both truths simultaneously: that technology can meaningfully contribute to food security, and that it cannot substitute for the trade policy, governance, and political will required to ensure that food reaches the people who need it. The Green Revolution of the twentieth century increased global food production dramatically and failed to prevent significant hunger, because the political and economic structures distributing that food were not adjusted to ensure access for the poorest. A second Green Revolution that makes the same mistake while also concentrating new technology capabilities in the hands of large commercial agricultural operations in wealthy countries would be a considerable missed opportunity.
The technologies most likely to close the food security gap for the most vulnerable are those that are accessible to smallholder farmers in low- and middle-income countries: low-cost soil sensors, mobile-accessible decision support tools, drought-tolerant open-source seed varieties, improved market access through digital platforms that reduce the power of intermediaries, and extension services that translate technical knowledge into practical on-farm decisions. These are less photogenic than robot strawberry pickers and satellite-guided tractors. They are the interventions that the evidence most clearly supports as likely to feed the people who need feeding.
Tomorrow we are leaving the fields and entering the ocean — specifically the emerging blue economy of ocean technology. Deep-sea mining, offshore wind at scale, ocean-based carbon capture, and the question of whether the last major unexploited frontier on Earth is about to be opened up in ways its governance is nowhere near ready for. See you then.
Switched On is a daily technology series covering the ideas, systems, and arguments shaping the digital world. Opinionated. Witty. Occasionally wrong. Always worth the argument.
