When crops rotted and farmers called for robots
In the summer of 2021, fields across Lincolnshire held crops that no one came to pick. Strawberries softened in the heat, brassicas bolted, and growers watched harvests worth thousands of pounds return to the soil. This was not a story about the future of technology — it was a story about a labour system that had quietly depended on EU free movement for decades and then, almost overnight, did not.
TED talks have long asked what automation means for human work. TEDx Grantham sits in the middle of the region where that question is being answered not in a conference hall but in actual fields, under actual economic pressure. Lincolnshire and the surrounding area grow roughly a third of England's vegetables. Food and farming account for 24% of Greater Lincolnshire's jobs — nearly double the 13% national average — and in South Holland district, agricultural employment exceeds 40% of the local workforce. Nowhere else in England has more at stake in the debate about farm robots.
The numbers from 2021 tell the story plainly. The government issued 30,000 seasonal worker visas against an industry-estimated need of 80,000 to 90,000. The gap was not closed by domestic recruits. It was not closed at all. Growers did not turn to robotics because automation was fashionable or because a startup arrived with a pitch deck. They turned to it because a structural crisis left them with few alternatives, and the machines that followed were, in important respects, a response to failure rather than a leap toward the future.
The gap between TED talks on automation and what Lincolnshire actually shows
TED talks on automation and the future of work tend to start from abundance: a founder's ambition, an efficiency gain waiting to be unlocked, a vision of workers freed from repetitive tasks to pursue something more creative. The implicit geography is usually a well-capitalised company that chose disruption. The timeline is elastic — change arrives when it is ready, and the people involved have some say in the pace.
Lincolnshire inverts nearly every one of those assumptions. No one here chose the timeline. The machines arrived because a labour system collapsed, not because a founder decided farming needed disrupting. That distinction shifts the TED conversation from 'how might automation change work?' to something sharper and considerably less comfortable: automation for whom, decided by whom, and absorbing what cost — when neither workers nor most growers set the terms?
The standard 'jobs of the future' framing also tends to treat agricultural employment as a modest share of a diversified economy, something that will transition as technology matures. A region where farming and the food chain dominate employment at the concentration found here makes that abstraction visible and urgent. The question of who gains and who is exposed cannot be deferred to a later slide. It is already unfolding across more than 50% of England's grade 1 agricultural land, in communities where the next harvest is not a thought experiment.
The research cluster behind the machines
Behind the Thorvald robots and Dyson's Carrington glasshouse sits a research infrastructure that was being built before the Brexit labour crisis made its outputs suddenly urgent.
In 2018, a £4.8 million Research England award brought together five universities — Lincoln, Cambridge, Reading, Hertfordshire, and East Anglia — under a programme called Ceres Agri-Tech. The collaboration has since produced four spin-out companies and 29 specialist jobs. One of those spin-outs, Fruitcast, uses AI to forecast crop yields — helping growers plan labour and logistics weeks ahead of harvest rather than guessing from experience alone. Another, Agaricus Robotics, is working on automated mushroom harvesting: a crop that has long resisted mechanisation because of how fast and unevenly mushrooms grow.
A further £4.9 million EPSRC grant supports LINCAM, which aims to establish the Lincolnshire–Cambridgeshire corridor as an agri-tech export hub. The underlying argument is that Lincolnshire's current pressures — labour scarcity, disease management, yield uncertainty — are problems other major agricultural regions will face next, and that solutions developed here can travel.
The people pipeline runs through AgriFoRwArdS, the world's first EPSRC Centre for Doctoral Training dedicated to agri-food robotics, run jointly by Lincoln, Cambridge, and East Anglia. Students begin with an MSc in robotics at Lincoln's Brayford Campus before moving into PhD research across partner sites. These are the researchers who translate what works in a controlled setting into something a working farm can actually deploy — the step between a proof of concept and a machine running through the night on a Lincolnshire strawberry crop.
UV-C robots working the night shift
Saga Robotics' Thorvald moves through strawberry rows after dark — not because the schedule is convenient, but because the biology demands it. UV-C light damages the DNA of fungal pathogens, but sunlight triggers a repair mechanism in those same organisms. Daytime application would undo much of what the treatment achieves. So Thorvald runs at night, and in doing so it also counts: each robot logs roughly 3 million data points per shift, recording flower and fruit numbers across every pass. A single machine is treating crops and building a yield forecast simultaneously.
Thorvald is now deployed across approximately 30% of UK tabletop strawberry production. It operates on conventional working farms — dropped into an existing growing environment rather than requiring the farm to be redesigned around it. That deployability matters: most Lincolnshire growers are not in a position to rebuild from scratch.
Dyson Farming's glasshouse at Carrington, by contrast, was designed with automation at its centre. The 26-acre site houses more than 1.2 million strawberry plants grown on rotating wheels — a configuration that brings the crop to the machine rather than routing the machine through a field. UV-C robots handle disease management; robotic arms assist with harvesting; AI decision-support software processes sensor data from throughout the structure. It is closer to a controlled manufacturing environment than a farm in any traditional sense.
Neither example is typical of the wider sector. Thorvald represents a technology that can be adopted incrementally; Dyson's glasshouse represents the capital-intensive end point that incremental adoption is pointing towards — but cannot reach without the scale and investment that most family farms in the region simply do not have.
The jobs that have gone are not the ones people assumed would go first
The common assumption about farm automation — that robots move first on the most numerous, most visible jobs — does not hold in Lincolnshire. Seasonal pickers arriving each summer to harvest strawberries, brassicas, and salad crops are not, for the most part, the workers whose roles have already changed. It is the people who walked the rows with spray equipment, or who spent hours scanning plants for early signs of disease, whose work the current generation of machines has absorbed.
That distinction matters because it complicates any clean account of who benefits and who loses. UV-C treatment and crop monitoring were skilled, time-sensitive tasks — chemical-handling certification and trained observation were required. Removing them from the human roster is a genuine structural change in farm employment, even if the harvest headcount looks largely the same.
The harvesting gap remains wide. Consistent expert assessment puts fully autonomous fruit-picking for delicate crops at five to ten years from widespread commercial viability. The dexterity required to pick a ripe strawberry without bruising it, at commercial speed, across variable field conditions, is a harder engineering problem than nighttime UV-C application. Home Office figures recorded 38,039 seasonal worker visas in the year to June 2025 — up 11% year-on-year — confirming that the human harvest workforce is very much still present. The UK Government's 2024 Seasonal Worker Visa review names automation as the long-term mechanism for reducing agricultural dependence on migrant labour, but the transition is measured in years, not seasons.
What that transition feels like from the inside is harder to establish. Workers in Lincolnshire's seasonal sector are predominantly migrants on short-term visas, with limited presence in English-language local media; growers, meanwhile, have little commercial incentive to publicise which roles a machine has taken over. The result is that the human dimension of this shift exists, for now, in aggregate statistics and policy documents rather than individual accounts.
What the next decade honestly looks like from here
The UK Government's seasonal worker scheme runs until 2029, with visa numbers set to taper as automation investment rises. This is a policy lever, not a market forecast — designed to make human pickers progressively harder to source, and machines progressively harder to defer. Lincolnshire growers have roughly five years to navigate that squeeze.
For operations at Dyson's scale, the direction is clear. For the smaller family farms that make up most of the region's agricultural sector, the arithmetic is considerably harder. Lincoln's research cluster has produced four spinouts and a pipeline of 40-plus innovations, but whether those innovations reach a price point accessible to a modest Fens holding is the central question that agri-tech optimism consistently sidesteps. Capital cost is where the gap between the sector's headline success stories and its everyday reality lives — and it is a gap the current evidence base does not yet quantify with any precision.
Thorvald is moving through Lincolnshire strawberry rows tonight, counting fruit in the dark while the policy clock runs toward 2029. The questions that TED talks tend to surface — who benefits from a transition, at what pace, and at whose expense — are being asked here, in these specific fields, without clean answers in sight. Robots are real, deployed, and growing in scope. Full automation of the seasonal harvest workforce is not imminent. What the human shape of the next decade looks like, for the smaller half of the region's farming economy, is still genuinely open.