The Drop 2026 on labour
In short
The jobs short of people are the hardest to hand to machines
At The Drop's twelve sessions on labour, investors and founders found people scarcest in fields, on building sites and in care, and AI most useful on a screen.
Last year, an agrifood investor said at the Ripple on farm robots, US field agriculture had 300,000 open positions and 800 applicants. Open fields have the biggest labour gap and the hardest engineering, since a field robot has to get past a log that fell from a tree the day before. The same mismatch recurred across The Drop's twelve sessions on labour. The work running out of people is the work hardest to hand to machines, and the work AI already does well happens on a screen.
The Drop framed the theme as two trends converging, fewer people entering the workforce and fewer jobs that need them. The sessions found them meeting in different places. Europe must fill 4.2 million construction jobs by 2035, a construction-robotics founder said, and 4.1 million of them replace people leaving the industry. In the Stockholm and Gothenburg regions there are half as many 20–25-year-olds as 30–35-year-olds, Greta Braun of Chalmers University of Technology said in a live podcast on care and AI. Malin Frithiofsson, the host, said the work most needed is the work least open to automation, physical and embodied care. AI cannot breastfeed in the middle of the night or play bingo with an elderly grandfather.
Screen work is where AI already saves on people. One group at the Ripple on AI and deep-tech timelines said teams now do as much as before with far fewer people, the biggest cost in building a company, though the physical validation step remains. At the Ripple on hardware after SaaS, a participant said what AI removes is mainly labour. The same Ripple heard that whoever builds the factories before robots can, perhaps a decade from now, will own them. Bio-manufacturing founders at the Ripple on hardware and biology called physical work their defence now that AI makes software easy to copy.
Buying the shop floor
When robots did sell, the customer was often buying an outcome. A start-up that builds solar farms with robots won its first four customers because data centres wanted two gigawatts in half the time first promised, and no human crew can work that fast, Raj Kapoor of Climactic said at the Ripple on strategy in physical AI. The contractors pay per panel placed. On farms, Julius Strauss of FoodLabs said, most founders now sell robotics as a service, and the farmer buys the promise that the field will be harvested. The start-ups still need people in the field when robots get stuck.
At the morning Ripple on Industry 4.0's productivity paradox, Jan Marchewski of Hitachi Ventures said incumbents are bad at architectural change because the shop floor mirrors how they make decisions and set budgets. Later sessions heard of start-ups that got round the incumbent by buying one. No excavation contractor would trial one company's electric autonomous dozer, so it bought a contractor with 100 million in revenue. A pre-seed investor's latest bet buys small injection-moulding firms outright, because changing a process needs ownership.
The other route asks the customer to change nothing. The timber-frame founder takes a builder's existing designs, charges less than the builder pays today and asks for no capital.
Where the saved hours go
Two stories about old machines, told hours apart, explained why the gains arrive late and where they go. Hillel Zand of Maniv said factories that swapped the central steam engine for one big electric motor saw no rise in productivity. The payoff came 25 to 30 years after the electric motor, when each machine got its own and the floor was rearranged. In the afternoon, Malin recalled that the washing machine was expected to free seven of the eight hours laundry took. Standards of cleanliness rose instead.
The morning treated missing gains as a problem. US manufacturing labour productivity grew an average of 3.4% a year between 1987 and 2007, Hillel said, and fell by an average of half a percent a year from 2010 to 2022. A participant who had spent years living in factories challenged the premise. What surface-mount machines delivered was lower cost, which customers constantly demand. "You don't really want to see that much productivity gain," they said.
The podcast asked who gets the gains when they come. Greta said people fill the hours AI saves with more tasks, and that productivity sometimes falls after companies invest because their people are not ready. Malin, playing devil's advocate, wondered whether that paradox is a saving grace, since companies driven by margins will cut headcount, not give staff time off, if demand does not follow. In Sweden, women again took more of the parental leave and the days off for sick children in 2025 and 2026, and researchers linked the shift to children falling ill less often. "If one person can do it, the woman does it," Malin said. A lighter load is what care AI promises.
Skills without a target
The batch also found skills going spare. At the Ripple on Europe's car industry, George Georgiadis of Evercurious VC asked about combustion-engine engineers in their 40s with 30 working years ahead. Several participants said Europe has the talent and lacks ambition and money willing to take risks. At the Ripple on where the green economy lives, Gosbert Chagula of Future Impact Ventures said the electric arc furnace replacing the blast furnace at Port Talbot cuts emissions by 90%, and 2,800 jobs go. A founder there said Germany's shrinking car industry holds know-how that hard-tech scale-ups need.
Caeden Noël, 18 and a founding member of the Helsinki residency FR8, made ambition the subject of an Expert Session. Society funnels its most capable young people into quant desks and consultancies, Caeden said, and Europe caps ambition to a number or a rule. Caeden believes ambition is learned, and grows with what you read and who is around you.
Climate start-ups compete with AI for the same people. At the Ripple on what the boldest will build, Anna Ottosson of Mudcake said a large AI company can pay five times what a start-up can. One founder had hired a former senior oil executive who took a 95% pay cut. An investor said the end goal must be big enough that a stake in a very profitable company outweighs an AI salary.
Ownership kept returning as the answer to who keeps the value once machines do more of the work. Malin described the other side. By spreading expertise, AI removes the scarcity that gives workers bargaining power over wages. In a recent Anthropic model of an extreme scenario for the US in 2030, which the host urged caution about, today's split of wealth, 60% to labour and 40% to capital, tips towards capital.
At the farm Ripple, every hand had gone up at the start for robots taking a much larger share of agtech by 2035. Asked why, after so much talk of difficulty, one participant gave a simple reason. "I think it's the plain fact that no one wants to be the farmer any longer." That is the case for the robots. Who will own them, and who will be paid while they do the work, the twelve sessions left open.