Over the past three years, Péter Fankhauser's industrial robots haven’t only been capable of climb stairs, but in addition jump between boxes, do backflips and perform other parkour-like tricks.
The robots weren’t programmed to perform these recent actions, but adapted to their environment using recent artificial intelligence models.
“These are the moments whenever you think, that is the subsequent revolution,” says Fankhauser, CEO of ANYbotics, a Zurich-based robotics start-up. “These things began to maneuver really artfully, and it's almost scary since the robots are fiddling with physics.”
Over the past decade, the $74 billion robotics sector has seen its capabilities increase attributable to significant advances in artificial intelligence, including advances in neural networks, systems that mimic the human brain.
The world's largest technology and AI corporations – including Google, OpenAI and Tesla – are racing to develop an AI “brain” that may autonomously control robots and transform entire industries from manufacturing to healthcare.
In particular, improved capabilities in computer vision and spatial reasoning have enabled robots to turn out to be more autonomous in navigating a wide selection of environments, from construction sites to grease rigs to city streets.
To train and program robots, engineers previously needed to hardwire rules and directions that taught the machine find out how to behave, often specific to every system or environment.
With the introduction of deep learning models lately, experts are capable of train AI software to make machines way more adaptable and aware of unexpected physical challenges in the true world and to learn independently.
Generative AI – a technology that may generate and analyze multimedia and text – has also enabled machines to raised understand the world around them and communicate more easily with humans. The technology has allowed people without programming skills to provide instructions to computers using text or voice instructions.
“It's like watching a toddler learn,” says Carina Namih, partner at Plural, a London-based early-stage investment fund. “Because the robots usually are not deterministically programmed but self-learning, the high development costs usually are not so high.”
While it is predicted that almost all advances can be seen in industrial environments and factory floors, large AI corporations are also increasingly specializing in human-like robots, so-called humanoids.
Earlier this 12 months, Google DeepMind announced numerous advances in its research, including using large language models to coach humanoid robots and help them higher and more safely understand and navigate their environment. World Labs, founded by the “godmother of AI” Fei-Fei Li, can also be working on this problem and has reached a $1 billion valuation in only 4 months.
ChatGPT developer OpenAI launched a robotics research group last month after dismantling its general-purpose robotics efforts in 2020 while investing in startups. These include Figure, which raised $675 million in February from investors including OpenAI, Microsoft, Jeff Bezos and Nvidia at a $2.6 billion valuation.
OpenAI also invested in Oslo-based 1X Robotics, which raised greater than $100 million this 12 months to develop on a regular basis bots for household tasks.
A recent evaluation by McKinsey estimates the worldwide marketplace for humanoid robots to be value just over a billion dollars, which is barely a fraction of your entire robot market. However, the market is growing by greater than 20 percent annually, 3 times faster than the marketplace for conventional industrial robots.
Experts warn that the technology remains to be inadequate and expensive. Chinese company Unitree Robotics sells its humanoid robot for $16,000. Elon Musk said Tesla will start using and manufacturing humanoid robots next 12 months and can sell them on a bigger scale from 2026.
Still, startups have tried to capitalize on the hype. According to data provider PitchBook, robotics and drone deals have reached $6.5 billion in value across 552 deals this 12 months, putting them heading in the right direction to surpass the $9.7 billion raised across 1,256 deals in all of 2023. However, total investment within the sector has been steadily declining since 2021.
“Robotics is a very difficult challenge – previous solutions were expensive and inflexible, which hindered adoption, especially amongst small and medium-sized corporations,” said Luciana Lixandru, a partner at enterprise capital firm Sequoia Capital, which has invested in numerous AI robotics startups, including RobCo and Collaborative Robotics. “Advances in AI may help overcome among the limitations.”
Much of the investment is in young corporations. Mytra, a robotics company for warehouse automation, announced this week that it has raised $78 million in three rounds. Munich-based RobCo raised $42.5 million in February to back its flexible robot hardware kits.
San Francisco-based Tetsuwan Scientific recently closed a $2.5 million funding round and goals to develop an AI robot scientist that may conduct research and physics experiments within the lab. CEO Cristian Ponce said robots could reproduce experiments with greater accuracy than humans, giving scientists more freedom for creativity and discovery.
“Generative AI has been applied to crazy things that don’t matter, like back-office tasks or accounting software, but applying generative AI to scientific discovery is essentially the most impactful thing we will do,” Ponce said.
At the identical time, the mass use of AI tools by consumers has had a domino effect on attitudes towards robotics, says Sonali Fenner of the management consultancy Slalom.
This has enabled corporations to contemplate using robots in customer-facing environments. Fenner cited the instance of a serious retail customer that deployed Spot, a robot dog from Boston Dynamics powered by Google's Gemini Pro model, in its stores to evaluate inventory.
“(The hype) has opened the door just a little bit by way of what is predicted in certain environments, and even in case you don't want Spot walking around your retail store, you might accept a rather less intrusive robot,” Fenner added.
Ahti Heinla, co-founder of Skype and CEO of delivery robot startup Starship Technologies, which has deployed the small grocery robots in over 100 cities and towns in Europe and the UK, said he was surprised at how easily people “perceive the robots as normal participants in public space and accept them as a matter after all”.
