Robots could soon be delivering your pizza An Estonian startup says its machines are now cheaper than human couriers June 11th 2026 If you live in San Francisco, you will often get a glimpse of the future— commuting in a self-driving taxi, say. In Milton Keynes? Not so much. But the English city, best known for its many roundabouts, is the place to go if you want to foresee a world without delivery drivers. It is one of the largest markets for Starship Technologies, an Estonian startup which claims to have cracked the problem of getting robots to deliver groceries more cheaply than people can. Designers of delivery robots face challenges familiar to anyone developing robotaxis. Starship has had to build a sensor array that its six-wheeled couriers, each the size of a beer cooler, can use to navigate along pavements
come rain or shine. That hardware must feed into an artificial-intelligence model that can pick the best route and carry on even if the connection to a data centre is lost. In some ways, though, delivery robots have it easy. With a 35kg robot travelling at 6kph (4mph) tops, safety is less of a worry than it is with a two- tonne car going at 110kph on a motorway. And a slightly bumpy ride won’t hurt a pizza. That said, whereas robotaxi firms often leave design to carmakers, robocouriers have no such option. Starship has by now created several generations of vehicles, and has optimised newer models for resilience and repairability. Gains can come from unexpected places, says Ahti Heinla, a co-founder. Starship’s latest batch of robots charge wirelessly, for instance, which is speedier and reduces wear on the charging ports. After 12 years of such improvements, Mr Heinla says, the cost of each delivery is now “significantly less” than that of paying a human to do it. Starship is aiming for a cost of less than £1 ($1.34) per delivery. “It’s not quite there yet, but not very far away,” Mr Heinla says. In 2018 the company had 127 robots, driving 116,000 kilometres in the year. By 2025 it had 2,414 robots, covering 5.2m kilometres. Along the way, it has reduced human interventions per kilometre by seven-eighths. Even so, at its scale rare problems, such as a robot failing in the middle of a street, add up. (The solution? A simple back-up computer designed purely to get it to the other side.) The little six-wheelers may soon change how cities look. But they are sure to irk some people, not least out-of-work delivery drivers. Then again, in Finland, Starship’s biggest market, the startup’s supermarket partner has had to urge sympathetic passers-by not to lift the robots out of snowdrifts when they get stuck—lest they hurt themselves while attempting a rescue. ■ To track the trends shaping commerce, industry and technology, sign up to “The Bottom Line”, our weekly subscriber-only newsletter on global business. This article was downloaded by zlibrary from https://www.economist.com//business/2026/06/07/robots-could-soon-be-delivering- your-pizza
Apple’s new Siri is a dark horse in the AI race The iPhone-maker does not need to build models to cash in on the technology June 11th 2026 Two years ago Apple announced its first foray into artificial intelligence. Built largely on in-house models, “Apple Intelligence” promised to turn the iPhone-maker’s boneheaded Siri assistant into a perspicacious PA as clever as the whizziest chatbots—with the added advantage of access to a user’s personal data, alongside various other superpowers. The effort was an embarrassing flop, with Apple having delivered little of what it set out to offer. Now it is taking a second bite. On June 8th, at its annual software jamboree, the company’s outgoing boss, Tim Cook, again unveiled a “new Siri”, which users can operate using their voice, a pull-down search bar or a chatbot-style
app. Rather than building its own models, the company is using ones made by Google, which competes at the frontier of AI. Apple is betting that its devices, and the personal data stored on them, will become the portals through which users get access to the technology. Will its strategy pay off? Apple’s earlier AI flubs have not caused it obvious harm. Its stock price is up by more than half over the past two years—less than for Google’s parent company, Alphabet, but more than for Amazon, Microsoft or Meta, all of which have torched stacks of cash in an effort to get ahead in the AI race. Apple, by contrast, has been able to sit back and take a cut of up to 30% of the revenues generated by chatbot apps installed on its devices. Even so, competition is looming. OpenAI, maker of Chatgpt, is working with Sir Jony Ive, lead designer for many of Apple’s most famous products, to make its own AI-powered gadget. Google and Meta are investing in smart glasses. And Amazon is rolling out new AI features on its home companion, Alexa (though at a recent demonstration your correspondent found it to be no less moronic than the old Siri). Apple has at least two big advantages over those hoping to use AI to dislodge the iPhone. First is the scope of its access. Many of Siri’s new skills rely on Apple’s ability to scan information such as a user’s messages and operate across apps. Second is Apple’s prowess in hardware and semiconductors. Many whizzy new features will run on the devices themselves, rather than needing to be routed through an external server, reducing delays and ensuring they can be used even without an internet connection. It will also mean that Apple will not need to invest in data centres to the same extent as other AI providers. (Some computationally demanding AI features that cannot be performed on devices, such as extending and reframing photos, will come with daily usage limits, though subscribers to Apple’s iCloud+ service may have more access.) Apple is reportedly paying Google $1bn a year for its technology—a pittance compared with what it would cost to develop an alternative in- house. And once users are hooked on Siri, Apple could conceivably switch out the underlying models, giving it the whip hand in negotiations. “It will be built on Google tech, but Apple’s going to own that relationship with the consumer,” points out Gil Luria of D.A. Davidson, an investment firm.
Investors, for their part, are still chewing over the announcement; Apple’s share price dropped by 2% on June 8th. That may reflect the fact that, after years of delay, the new features are still not ready for consumers: the upgraded Siri will be available in America in autumn, but not on iPhones in the European Union or any Apple devices in China owing to regulatory snags. The new Siri will also not work in languages other than English at first. John Ternus, Apple’s incoming boss and current hardware supremo, did not speak at the conference, which focused only on software. But the company’s leisurely timeline means he will oversee most of the rollout. Horace Dediu, a veteran Apple analyst, points out that, although the company works slowly, “it tends to deliver eventually.” Mr Ternus will have to prove that is still true. ■ To track the trends shaping commerce, industry and technology, sign up to “The Bottom Line”, our weekly subscriber-only newsletter on global business. This article was downloaded by zlibrary from https://www.economist.com//business/2026/06/09/apples-new-siri-is-a-dark-horse-in- the-ai-race
Another new boss aims to fix the world’s biggest chocolate-maker Will Hein Schumacher’s turnaround plan for Barry Callebaut work? June 11th 2026 ITS products tempt you from supermarket shelves everywhere, but you may not know its name. Formed 30 years ago when Klaus Jacobs, a German-born billionaire, folded together France’s Cacao Barry and Belgium’s Callebaut, Barry Callebaut is the world’s biggest chocolate-maker. The Zurich-based firm buys about a quarter of the annual global cocoa harvest and turns it into chocolate for Magnum’s ice creams, Nestlé’s KitKats and Mondelez’s Cadbury and Milka brands. Jacobs died in 2008. His family remain the biggest shareholders, with a stake of around a third, but lately this has brought them little joy. Barry Callebaut has struggled for years with soft sales and high debt. The share
price has melted; it is half what it was five years ago. Many of the firm’s woes were self-inflicted, but the whole industry was hammered in 2023-24 by crop failures related to climate change that drove cocoa prices to alarming heights. They leapt from around $2,500 a tonne to peak at $12,000 in 2024. (They are now around $4,000.) In January Barry Callebaut appointed Hein Schumacher, a Dutchman and former boss of Unilever, as its fourth chief executive in six years. In April he said that operating profit in the year to August would fall by a “mid-teens” percentage, rather than rise as previously projected, as the company prioritised rebuilding sales volumes over boosting margins. The share price dropped by 16% on the day. Can Mr Schumacher revitalise Barry Callebaut? On June 2nd he served up a plan. In the medium term he wants sales volumes to grow by 2-4% a year. That would be low by historical standards, but a welcome change after years of shrinkage: in the previous financial year, ending in August 2025, they fell by 6.8%. Mr Schumacher will concentrate on ten growth markets, including Brazil and Indonesia. He will also prioritise Gourmet, a business supplying top-quality chocolate to master chocolatiers, pastry chefs and restaurants, and the similarly profitable “speciality” segment, which, for instance, makes chocolate that resists melting in hot climates. Analysts’ early verdicts are mixed. Bank Vontobel, a Swiss private bank, calls the plan a “crucial step in restoring growth”. Mr Schumacher “is spending a lot of time visiting the firm’s seven biggest clients and listening to their needs”, notes Vontobel’s Matteo Lindauer. But analysts at Helvetische Bank, another Swiss financial institution, are more sceptical: Barry Callebaut will evidently remain in turnaround mode for years, they write. Restoring investors’ appetites will not be easy. In February Nicolas Jacobs, a board member and co-chairman of the family holding company, sold shares worth around SFr14m. That is a mere shaving off a chunky bar—Barry Callebaut’s market capitalisation is SFr6.1bn. Some analysts, however, took the sale as a sign that the family felt recovery would be slow at best.