Robots Hit the Streets as Demand for Food Delivery Grows

A food delivery robot crosses a street in Ann Arbor, Mich. on Thursday, Oct. 7, 2021. (AP Photo/Carlos Osorio)
A food delivery robot crosses a street in Ann Arbor, Mich. on Thursday, Oct. 7, 2021. (AP Photo/Carlos Osorio)
TT

Robots Hit the Streets as Demand for Food Delivery Grows

A food delivery robot crosses a street in Ann Arbor, Mich. on Thursday, Oct. 7, 2021. (AP Photo/Carlos Osorio)
A food delivery robot crosses a street in Ann Arbor, Mich. on Thursday, Oct. 7, 2021. (AP Photo/Carlos Osorio)

Robot food delivery is no longer the stuff of science fiction. But you may not see it in your neighborhood anytime soon.

Hundreds of little robots __ knee-high and able to hold around four large pizzas __ are now navigating college campuses and even some city sidewalks in the US, the UK and elsewhere. While robots were being tested in limited numbers before the coronavirus hit, the companies building them say pandemic-related labor shortages and a growing preference for contactless delivery have accelerated their deployment.

“We saw demand for robot usage just go through the ceiling,” said Alastair Westgarth, the CEO of Starship Technologies, which recently completed its 2 millionth delivery. “I think demand was always there, but it was brought forward by the pandemic effect."

Starship has more than 1,000 robots in its fleet, up from just 250 in 2019. Hundreds more will be deployed soon. They're delivering food on 20 US campuses; 25 more will be added soon. They're also operating on sidewalks in Milton Keynes, England; Modesto, California; and the company’s hometown of Tallin, Estonia.

Robot designs vary; some have four wheels and some have six, for example. But generally, they use cameras, sensors, GPS and sometimes laser scanners to navigate sidewalks and even cross streets autonomously. They move around 5 mph.

Remote operators keep tabs on multiple robots at a time but they say they rarely need to hit the brakes or steer around an obstacle. When a robot arrives at its destination, customers type a code into their phones to open the lid and retrieve their food.

The robots have drawbacks that limit their usefulness for now. They’re electric, so they must recharge regularly. They're slow, and they generally stay within a small, pre-mapped radius.

They’re also inflexible. A customer can’t tell a robot to leave the food outside the door, for example. And some big cities with crowded sidewalks, like New York, Beijing and San Francisco, aren’t welcoming them.

But Bill Ray, an analyst with the consulting firm Gartner, says the robots make a lot of sense on corporate or college campuses, or in newer communities with wide sidewalks.

“In the places where you can deploy it, robot delivery will grow very quickly,” Ray said.

Ray said there have been few reports of problems with the robots, other than an occasional gaggle of kids who surround one and try to confuse it. Starship briefly halted service at the University of Pittsburgh in 2019 after a wheelchair user said a robot blocked her access to a ramp. But the university said deliveries resumed once Starship addressed the issue.

Patrick Sheck, a junior at Bowling Green State University in Bowling Green, Ohio, gets deliveries from a Starship robot three or four times a week as he’s leaving class.
“The robot pulls up just in time for me to get some lunch,” Sheck said. Bowling Green and Starship charge $1.99 plus a service fee for each robot delivery.

Rival Kiwibot, with headquarters in Los Angeles and Medellin, Columbia, says it now has 400 robots making deliveries on college campuses and in downtown Miami.

Delivery companies are also jumping into the market. Grubhub recently partnered with Russian robot maker Yandex to deploy 50 robots on the campus of Ohio State University in Columbus, Ohio. Grubhub plans to add more campuses soon, although the company stresses that the service won’t go beyond colleges for now.

US delivery orders jumped 66% in the year ending in June, according to NPD, a data and consulting firm. And delivery demand could remain elevated even after the pandemic eases because customers have gotten used to the convenience.

Ji Hye Kim, chef and managing partner of the Ann Arbor, Michigan, restaurant Miss Kim, relied heavily on robot delivery when her dining room was closed last year. Kim had partnered with a local robot company, Refraction AI, shortly before the pandemic began.

Kim prefers robots to third-party delivery companies like DoorDash, which charge significantly more and sometimes cancel orders if they didn’t have enough drivers. Delivery companies also bundle multiple orders per trip, she said, so food sometimes arrives cold. Robots take just one order at a time.

Kim said the robots also excite customers, who often post videos of their interactions.

“It’s very cute and novel, and it didn’t have to come face to face with people. It was a comfort,” Kim said. Delivery demand has dropped off since her dining room reopened, but robots still deliver around 10 orders per day.

While Kim managed to hang on to her staff throughout the pandemic, other restaurants are struggling to find workers. In a recent survey, 75% of US. restaurant owners told the National Restaurant Association that recruiting and retaining employees is their biggest challenge.

That has many restaurants looking to fill the void with robot delivery.

“There is no store in the country right now with enough delivery drivers,” said Dennis Maloney, senior vice president and chief digital officer at Domino’s Pizza.

Domino’s is partnering with Nuro, a California startup whose 6-foot-tall self-driving pods go at a maximum speed of 25 mph on streets, not sidewalks. Nuro is testing grocery and food delivery in Houston, Phoenix and Mountain View, California.

Maloney said it's not a question of if, but of when, robots will start doing more deliveries. He thinks companies like Domino's will eventually use a mix of robots and drivers depending on location. Sidewalk robots could work on a military base, for example, while Nuro is ideal for suburbs. Highway driving would be left to human workers.

Maloney said Nuro delivery is more expensive than using human drivers for now, but as the technology scales up and gets more refined, the costs will go down.

For cheaper sidewalk robots __ which cost an estimated $5,000 or less __ it's even easier to undercut human delivery costs. The average Grubhub driver in Ohio makes $47,650 per year, according to the job site Indeed.com.

But robots don’t always cost delivery jobs. In some cases, they help create them. Before Starship’s robots arrived, Bowling Green didn’t offer delivery from campus dining spots. Since then, it has hired more than 30 people to serve as runners between kitchens and robots, Bowling Green dining spokesman Jon Zachrich said.

Brendan Witcher, a technology analyst with the consulting firm Forrester, says it’s easy to get excited about the Jetsons-like possibility of robot delivery. But ultimately, robots will have to prove they create an advantage in some way.

“It’s possible that we see this emerge into something else,” he said. “But it’s the right time and place for companies considering robots to test them and learn from them and do their own evaluation.”



Meta Shares Skyrocket, Microsoft Slides on Wall Street after Earnings

A Microsoft logo is seen a day after Microsoft Corp's $26.2 billion purchase of LinkedIn Corp, in Los Angeles, California, US, June 14, 2016. REUTERS/Lucy Nicholson
A Microsoft logo is seen a day after Microsoft Corp's $26.2 billion purchase of LinkedIn Corp, in Los Angeles, California, US, June 14, 2016. REUTERS/Lucy Nicholson
TT

Meta Shares Skyrocket, Microsoft Slides on Wall Street after Earnings

A Microsoft logo is seen a day after Microsoft Corp's $26.2 billion purchase of LinkedIn Corp, in Los Angeles, California, US, June 14, 2016. REUTERS/Lucy Nicholson
A Microsoft logo is seen a day after Microsoft Corp's $26.2 billion purchase of LinkedIn Corp, in Los Angeles, California, US, June 14, 2016. REUTERS/Lucy Nicholson

Shares in Meta skyrocketed by 10 percent at opening on Wall Street on Thursday, a day after the social media giant posted better than expected earnings as the company invests heavily in artificial intelligence.

Microsoft, whose earnings disappointed analysts, saw its share price tumble by 10 percent, with investors showing concern for the return on investment for the software giant's spending on AI.


Samsung Logs Best-ever Profit on AI Chip Demand

South Korean tech giant Samsung Electronics posted record quarterly profits on Thursday, riding strong market demand for its artificial intelligence chips. Jung Yeon-je / AFP/File
South Korean tech giant Samsung Electronics posted record quarterly profits on Thursday, riding strong market demand for its artificial intelligence chips. Jung Yeon-je / AFP/File
TT

Samsung Logs Best-ever Profit on AI Chip Demand

South Korean tech giant Samsung Electronics posted record quarterly profits on Thursday, riding strong market demand for its artificial intelligence chips. Jung Yeon-je / AFP/File
South Korean tech giant Samsung Electronics posted record quarterly profits on Thursday, riding strong market demand for its artificial intelligence chips. Jung Yeon-je / AFP/File

South Korean tech giant Samsung Electronics posted record quarterly profits Thursday, riding massive market demand for the memory chips that power artificial intelligence.

A global frenzy to build AI data centers and develop the fast-evolving technology has sent orders for advanced high bandwidth memory microchips soaring.

That is also pushing up prices for less flashy chips used in consumer electronics -- threatening higher prices for phones, laptops and other devices worldwide.

In the quarter to December 2025, Samsung said it saw "its highest-ever quarterly consolidated revenue at KRW 93.8 trillion (US$65.5 billion)", a quarter-on-quarter increase of nine percent.

"Operating profit was also an all-time high, at KRW 20.1 trillion," the company said.

The dazzling earnings came a day after a key competitor, South Korean chip giant SK hynix, said operating profit had doubled last year to a record high, also buoyed by the AI boom.

The South Korean government has pledged to become one of the top three AI powers, behind the United States and China, with Samsung and SK hynix among the leading producers of high-performance memory.

Samsung said Thursday it expects "AI and server demand to continue increasing, leading to more opportunities for structural growth".

Annual revenue stood at 333.6 trillion won, while operating profit came in at 43.6 trillion won. Sales for the division that oversees its semiconductor business rose 33 percent quarter-on-quarter.

The company pointed to a $33.2 billion investment in chip production facilities -- pledging to continue spending in "transitioning to advanced manufacturing processes and upgrading existing production lines to meet rising demand".

- 'Clearly back' -

Major electronics manufacturers and industry analysts have warned that chipmakers focusing on AI sales will cause higher retail prices for consumer products across the board.

This week US chip firm Micron said it was building a $24 billion plant in Singapore in response to AI-driven demand that has caused a global shortage of memory components.

SK hynix announced Wednesday that its operating profit had doubled last year to a record 47.2 trillion won.

The company's shares have surged some 220 percent over the past six months, while Samsung Electronics has risen about 130 percent, part of a huge global tech rally fueled by optimism over AI.

Both companies are on the cusp of producing next-generation high-bandwidth "HBM4" chips for AI data centers, with Samsung reportedly due to start making them in February.

American chip giant Nvidia -- now the world's most valuable company -- is expected to be one of Samsung's customers for HBM4 chips.

But Nvidia has reportedly allocated around 70 percent of its HBM4 demand to SK hynix for 2026, up from the market's previous estimate of 50 percent.

"Samsung is clearly back and we are expecting them to show a significant turnaround with HBM4 for Nvidia's new products -- helping them move past last year's quality issues," Hwang Min-seong, research director at market analysis firm Counterpoint, told AFP.

But SK still "maintains a market lead in both quality and supply" of a number of key components, including Dynamic Random Access Memory chips used in AI servers, he said.

SK also this week said it will set up an "AI solutions firm" in the United States, committing $10 billion and weighing investments in US companies.


Google Unveils AI Tool Probing Mysteries of Human Genome

A Google logo is seen at a company research facility in Mountain View, California, US, May 13, 2025. (Reuters)
A Google logo is seen at a company research facility in Mountain View, California, US, May 13, 2025. (Reuters)
TT

Google Unveils AI Tool Probing Mysteries of Human Genome

A Google logo is seen at a company research facility in Mountain View, California, US, May 13, 2025. (Reuters)
A Google logo is seen at a company research facility in Mountain View, California, US, May 13, 2025. (Reuters)

Google unveiled an artificial intelligence tool Wednesday that its scientists said would help unravel the mysteries of the human genome -- and could one day lead to new treatments for diseases.

The deep learning model AlphaGenome was hailed by outside researchers as a "breakthrough" that would let scientists study and even simulate the roots of difficult-to-treat genetic diseases.

While the first complete map of the human genome in 2003 "gave us the book of life, reading it remained a challenge", Pushmeet Kohli, vice president of research at Google DeepMind, told journalists.

"We have the text," he said, which is a sequence of three billion nucleotide pairs represented by the letters A, T, C and G that make up DNA.

However, "understanding the grammar of this genome -- what is encoded in our DNA and how it governs life -- is the next critical frontier for research," said Kohli, co-author of a new study in the journal Nature.

Only around two percent of our DNA contains instructions for making proteins, which are the molecules that build and run the body.

The other 98 percent was long dismissed as "junk DNA" as scientists struggled to understand what it was for.

However, this "non-coding DNA" is now believed to act like a conductor, directing how genetic information works in each of our cells.

These sequences also contain many variants that have been associated with diseases. It is these sequences that AlphaGenome is aiming to understand.

- A million letters -

The project is just one part of Google's AI-powered scientific work, which also includes AlphaFold, the winner of 2024's chemistry Nobel.

AlphaGenome's model was trained on data from public projects that measured non-coding DNA across hundreds of different cell and tissue types in humans and mice.

The tool is able to analyze long DNA sequences then predict how each nucleotide pair will influence different biological processes within the cell.

This includes whether genes start and stop and how much RNA -- molecules which transmit genetic instructions inside cells -- is produced.

Other models already exist that have a similar aim. However, they have to compromise, either by analyzing far shorter DNA sequences or decreasing how detailed their predictions are, known as resolution.

DeepMind scientist and lead study author Ziga Avsec said that long sequences -- up to a million DNA letters long -- were "required to understand the full regulatory environment of a single gene".

And the high resolution of the model allows scientists to study the impact of genetic variants by comparing the differences between mutated and non-mutated sequences.

"AlphaGenome can accelerate our understanding of the genome by helping to map where the functional elements are and what their roles are on a molecular level," study co-author Natasha Latysheva said.

The model has already been tested by 3,000 scientists across 160 countries and is open for anyone to use for non-commercial reasons, Google said.

"We hope researchers will extend it with more data," Kohli added.

- 'Breakthrough' -

Ben Lehner, a researcher at Cambridge University who was not involved in developing AlphaGenome but did test it, said the model "does indeed perform very well".

"Identifying the precise differences in our genomes that make us more or less likely to develop thousands of diseases is a key step towards developing better therapeutics," he explained.

However, AlphaGenome "is far from perfect and there is still a lot of work to do", he added.

"AI models are only as good as the data used to train them" and the existing data is not very suitable, he said.

Robert Goldstone, head of genomics at the UK's Francis Crick Institute, cautioned that AlphaGenome was "not a magic bullet for all biological questions".

This was partly because "gene expression is influenced by complex environmental factors that the model cannot see", he said.

However, the tool still represented a "breakthrough" that would allow scientists to "study and simulate the genetic roots of complex disease", Goldstone added.