Eclipse; Lightspeed Venture Partners; Winni Wintermeyer Photography
- Investors are piling into robotics as AI moves from screens into the real world.
- Robotics and physical AI startups have raised more than $23 billion this year.
- BI identified 22 investors shaping the boom, from humanoids to defense tech.
For decades, software ate the world. Venture investors chased companies that could grow without factories, supply chains, or machines. Hardware was almost a dirty word.
But now, Silicon Valley is getting physical.
As the software boom cools and AI moves from chatbots into the physical world, investors are piling into robotics and so-called "physical AI," a term popularized by Nvidia CEO Jensen Huang to describe AI systems that can act in the real world. The pitch is that if AI transformed what happens on screens, robotics could transform work in warehouses, hospitals, construction sites, homes, and battlefields.
Venture capital investment in global robotics and physical AI has grown from around $4 billion in 2019 to $26 billion in 2025, according to PitchBook data. This year, companies in the space have raised more than $23 billion.
Several forces are driving the boom. Sensors, cameras, actuators, and other components have become cheaper and more capable. AI allows robots to move beyond rigid, preprogrammed behaviors. Talent from Tesla, Waymo, Amazon, and other hardware-building companies has learned how to deploy technology in the real world. And labor shortages and geopolitical pressure to rebuild supply chains have made automation more urgent.
"Now the cool kids have arrived," said Matt Ocko, cofounder and managing partner at DCVC, a venture firm known for backing deep-tech companies. Ocko said he and fellow investor Steve Jurvetson, an early backer of SpaceX, coined the term "deep tech" decades ago to describe companies "delivering existentially necessary and valuable results in the physical world."
The rush has also brought in investors with little experience in the category. Industry veterans say "hardware tourists" are flooding robotics cap tables, drawn in by hype and underestimating how hard it is to build machines that work reliably outside a demo.
Business Insider identified 22 investors, from established names to rising stars, who are shaping the robotics boom. Their bets span humanoids, autonomous vehicles, warehouse automation, defense robotics, and the software models that could define the next generation of machines. The list is alphabetical.
Eclipse
Madigan-Curtis began her Silicon Valley career at Apple, where she worked on global operations for the first Apple Watch. She later joined Samsara as an early executive and led operations as the company scaled.
When Eclipse approached her about investing, Madigan-Curtis saw an opportunity to back companies bringing technology into the physical world.
"It felt to me like there was such a gap between the digital world and the real world we all live in — the one where we drive down roads and go to the doctor," Madigan-Curtis said. "I thought, when does technology start to penetrate the real world?"
At Eclipse, Madigan-Curtis is part of a team of former operators investing in physical-world companies. She led the firm's investment in Verkada, which builds security cameras and software for businesses, and helped incubate Bedrock Robotics, a startup founded by former Waymo engineers that is automating heavy construction equipment. She also invested in Simbe Robotics, whose Tally robot scans store shelves to help retailers track inventory and pricing.
"For a long time, the hardware was the hardest part about robotics," Madigan-Curtis said. "The real frontier now is the intelligence layer and building true embodied AI."
Bain Capital Ventures
Agarwal spotted the promise of warehouse robots long before Amazon wanted in. In 2004, he backed Kiva Systems, which pioneered warehouse automation by using fleets of mobile robots to bring products directly to human workers. Agarwal led every successive funding round, and when Amazon bought Kiva in 2012 for $775 million, BCV was the company's only institutional investor.
More recent bets include Vention, which lets manufacturers design and order custom factory equipment online; Gather AI, which uses autonomous drones to track inventory inside warehouses; and Mind Robotics, a Rivian spinout building AI-powered industrial robots.
Agarwal is perhaps the most prominent robotics investor arguing against humanoids. In a Wall Street Journal op-ed last year, he wrote that humanoids would prove to be a "parlor trick" with few practical uses.
His portfolio reflects that thesis: Gather AI's drones fly through warehouses, and Mind Robotics is focused on automating factory tasks rather than building human-shaped machines.
"There's a reason why humans fly planes and drive in cars," Agarwal said. "Because wheels and wings are more efficient than walking."
Lumafield
Zuberi began investing in deep tech nearly two decades ago, when a small cohort of investors, including him, realized most VCs were ignoring entire industries, including energy, defense, and 3D printing.
"We realized there's a very big TAM [total addressable market] that Silicon Valley was just not addressing," Zuberi said.
After more than a decade at Lux Capital, where he backed companies including Applied Intuition, which builds software for autonomous vehicles, and Saildrone, which makes autonomous ocean drones, Zuberi struck out on his own last year. Red Glass Ventures, named after Zuberi's signature glasses, is an early-stage fund where he works closely with founders and early teams.
Zuberi's robotics investments include a foundation model startup, a robotics data training startup, and Foundry Robotics, his only public robotics investment so far. Foundry is developing robots to automate factory assembly.
At the same time, Zuberi is wary of the hype around robotics.
"My yellow flags are up. The space is overheated, overcrowded, and a bunch of noisy investments are being made. People will lose money. We are somewhere near the top of the hype curve," he said.
Carla D.A.
Zhan did robotics research in college but became frustrated with the field's slow pace. He shifted into database research and joined Facebook after graduating.
He returned to robotics as an investor.
"Everyone around me agrees that it's too late to start a company in coding. Robotics is going to be the next major category," Zhan said.
Zhan practices what he describes as "nerdy investing." At CRV, he invested in Dyna Robotics, which is building foundation models for robots, and Skild AI, which is developing a general-purpose "brain" meant to work across different kinds of robots and tasks. On Skild, he missed the seed round, which went to his sister, Sequoia partner Stephanie Zhan, and Lightspeed's Raviraj Jain. He later joined the Series A.
Zhan's investments reflect his belief that robotics is still in the stage where the basic building blocks are being created. Instead of every robotics company teaching its machines from scratch, he believes companies like Dyna and Skild will provide the AI models that many future robotics companies use to help their robots see, understand instructions, and decide how to move.
Bessemer Venture Partners
Levine and Goldberg are helping lead Bessemer's robotics and physical AI investment strategy. The two have argued that robotics is nearing its "GPT-2.5 moment": models are getting better, but the gap between demos and real-world deployment remains wide.
The firm has backed a range of robotics companies, including Waymo, the autonomous-driving company; Mind Robotics, which is building factory robots; Foxglove, a robotics software company; Breaker, a defense robotics startup; DroneDeploy, which uses drones and cameras to map worksites; Auterion, which builds software for drones; and ANYbotics, which makes four-legged robots for industrial inspection.
As their portfolio reflects, Levine and Goldberg are especially bullish on defense robotics, predicting the category will produce the first $50 billion-plus robotics IPOs. They reject the idea that robotics is in a bubble, arguing that too little capital is flowing into the sector given the size of the opportunity.
"There will be 100,000x more robots on Earth in the next 10-20 years," Levine wrote in a recent Bessemer report.
Barbara Kinney
Lichter joined Emerson Collective in 2021 and leads the firm's physical AI investments.
Her portfolio includes Field AI, a startup founded by former NASA Jet Propulsion Laboratory researchers. It builds software to help robots move through challenging environments, such as construction sites, on their own.
"Field combines three things we look for: a world-class technical team, a differentiated approach to a hard problem, and early evidence that the technology is working," Lichter said.
Her other robotics bets include Physical Intelligence, which builds general-purpose AI models for robots, and Agrippa, which applies robotics to agriculture.
For decades, Lichter argues, robots were too rigid to be broadly useful. Now, she sees signs that robots are beginning to handle new environments and tasks they weren't explicitly built for, though most of that progress remains in research labs.
"The physical transformation is just getting started," Lichter said. "And the prize is vastly larger than most people realize."
Todd Tankersley
Shah has been drawn to robotics since college, when she studied engineering at Carnegie Mellon University. At Obvious Ventures, Shah invests around a simple thesis: automate work that is dull, dirty, or dangerous. The firm has backed Dexterity, a warehouse robotics startup, and Pyka, which builds autonomous electric aircraft.
Shah's focus is general-purpose robotics, and she's currently interested in dexterous manipulation, the challenge of teaching robots to use their hands with human-like skill. The "hands problem," as it is often called in robotics, remains one of the field's hardest bottlenecks. That led to her investment in Eka Robotics, cofounded by MIT professor Pulkit Agrawal and former Google DeepMind robotics researcher Tuomas Haarnoja.
"I was looking for a world-class research team that's also commercially-minded," Shah said.
Shah said Eka's early demos, which show robot grippers nimbly clasping berries without squishing them and quickly screwing in lightbulbs, suggest its approach could help robots develop more capable hands over time. That could open the door to uses beyond factories and warehouses, including restaurants and homes.
She has also written personal angel checks into Field AI and XDOF, which develops infrastructure for robotics data collection and training.
Kris Connor/Getty Images
Boyle's robotics investments are rooted in Andreessen Horowitz's American Dynamism practice, which she cofounded. The strategy focuses on companies supporting the national interest across aerospace, defense, manufacturing, and critical infrastructure.
Before joining Andreessen Horowitz, Boyle was a reporter at The Washington Post and later a partner at General Catalyst, where she invested early in Anduril, the defense tech startup building autonomous systems for the military.
She joined Andreessen Horowitz in 2022. Her portfolio includes Hadrian, which automates precision manufacturing for aerospace and defense; Saronic, which builds autonomous ships for the military; and Castelion, a defense startup developing hypersonic weapons.
Techarena
Chen began her career on Wall Street. While running a mortgage-trading desk at Barclays, she started angel investing, including in an early satellite-constellation company. "I realized that I wanted to spend my career investing in this space," Chen said.
She later joined DCVC, where she invested in Agility Robotics, the humanoid robot maker; Fulfil, which builds automated grocery-fulfillment systems; and Recycleye, a robotics company using AI to sort waste for recycling facilities. Chen then became a founding partner at the NATO Innovation Fund, a venture fund backed by 24 NATO allies that is investing more than $1 billion in deep tech.
"My overall approach on the investing side has always been the same, which is how do we get to unit economics? What are the companies and industries where that is possible, or will soon be possible?" Chen said.
Her focus has shifted as hardware, computing power, and simulation tools have improved. She is now launching a new fund backing physical AI companies that address the economy's hardest physical bottlenecks: supply chains, labor, industrial resilience, and critical infrastructure.
Emergence Capital
Before becoming a VC, Saper developed large solar projects in India. To get aerial images of remote sites in Rajasthan's Thar Desert, he once had to hire a hot-air balloonist to take photos.
So when he later met DroneDeploy as a young investor, the use case was obvious. The company's software lets drones map and photograph sites for industries like construction, agriculture, and real estate. Emergence backed DroneDeploy in 2015, with Kevin Spain leading the deal.
"Being a physical-world AI investor benefits a lot from having worked in the environments that you are funding," Saper said.
A decade later, Saper and Spain co-led Emergence's investment in Bedrock Robotics, a startup founded by former Waymo engineers that's automating heavy construction equipment like excavators. The problem was familiar to Saper, who had seen how hard it can be to find skilled labor on construction sites.
Spain has also backed Physical Intelligence, which is building general-purpose AI models for robots.
"We believe that there are going to be foundation models that any robotics company can pick up and attach to their hardware to facilitate any kind of physical-world task, and Physical Intelligence is building that model," Spain said.
Winni Wintermeyer Photography
Ocko is a self-described "stubborn nerd" who has spent the past 25 years backing deep tech companies: startups built on scientific breakthroughs and advanced engineering rather than software alone.
On the day we spoke, the Trump administration awarded $2 billion in grants to nine quantum-computing companies, including $100 million each to DCVC-backed Atom Computing and Rigetti Computing. To Ocko, who founded DCVC in 2011, it was proof that deep tech bets can take years to mature, then become strategically essential and wildly lucrative.
DCVC's robotics portfolio includes nearly two dozen companies, including humanoid startup Agility Robotics, autonomous loading startup Slip Robotics, automated grocery fulfillment company Fulfil, and autonomous construction equipment company AIM.
"These companies are not the subject of week-by-week fawning Forbes articles," Ocko said. "But when they win, they win asymptotically. They win gigantically."
Hieu Tran
Mehta is known for concentrated, long-term bets on technology companies. Greenoaks began paying close attention to robotics more than five years ago, as AI and computer vision started making machines more useful in the real world.
Greenoaks has backed The Bot Company, the home-robot startup founded by Kyle Vogt, who previously sold Twitch and Cruise. The startup is building a wheeled household robot, rather than a humanoid, designed to help with everyday tasks. Greenoaks led a $150 million round in 2025 that valued the company at $2 billion.
Greenoaks led a Series B investment in Mytra, a company building warehouse robots that move goods across a modular 3-D storage system without requiring customers to rebuild their facilities. The firm has also invested in Mind Robotics, a Rivian spinout building robots for industrial deployment, and Physical Intelligence, which is building general-purpose AI models for the physical world.
Greenoaks' investments have been characterized by close partnership with founding teams and a focus on exceptional customer experiences.
Lightspeed Venture Partners
Jain's interest in robotics began when he moved from a small town in India to Mumbai for college and joined IIT Bombay's robotics club. Years later, at Lightspeed, he started looking for companies using computer vision to bring machines into the physical world.
He first backed Dexterity, a warehouse robotics startup. Later, after what he described as many "near-death experiences" testing self-driving cars, he invested in Aurora, an autonomous-trucking company.
Jain's latest robotics bet is Skild AI, where he was the founding investor and now serves on the board. The startup is building a general-purpose "brain" meant to work across different kinds of robots and tasks. Jain's thesis is that general-purpose robots will not arrive all at once because there still isn't enough real-world data to train them.
"My view is that we need to develop specific skills in robotics and build those into economically viable models, which will then converge into truly general-purpose robotics," Jain said.
Conviction
Guo spent nearly a decade at Greylock, becoming the firm's youngest general partner in 2018, before leaving in 2022 to launch Conviction, an early-stage AI fund.
She was the first backer of Sunday Robotics in 2024, a home-robot startup she said stood out because of its mix of "completely cracked researchers, a real commitment to building products that got deployed, and an Elon-style religion on delivering a mass-consumer product."
Sunday is preparing to launch Memo this fall as part of a beta program. Memo is a general-purpose wheeled home robot that can do dishes, fold laundry, and pull an espresso shot.
Guo is clear-eyed about the difficulty of the category. Robotics is capital-intensive and technically demanding, and American companies have struggled in adjacent hardware races, from self-driving cars to consumer electronics. Chinese manufacturers are formidable competitors on hardware costs. But Guo said robotics also requires a complete system combining hardware, software, and AI, plus top-tier research talent.
She is betting on the space because she believes the US needs robots and has the talent to build general-purpose systems.
"In terms of AI that will change people's lives, I can think of few things as valuable as giving people their time back, or filling some of the gaps we have in labor, such as in health and elder care," Guo said.
Eclipse
Winterroth is part of Eclipse's deep bench of investors focused on the physical economy. Since joining the firm at its founding in 2015, he said pure software investments have never excited him. "I've always been someone who enjoys zigging when other people are zagging," Winterroth said.
One early win was 6 River Systems, the warehouse robotics startup Shopify acquired for $450 million in 2019. The deal gave Winterroth a firsthand view of what it takes to build and sell robotics products.
Since then, he has built a broad robotics portfolio. He invested in Wayve in 2019, when the UK autonomous-driving startup's AI-first approach was still considered contrarian and the team had fewer than 20 people working out of a small Cambridge garage. Wayve is now valued at $8.6 billion and has a deal with Uber to launch self-driving vehicles in more than 10 markets.
Winterroth is also an investor and board member at Mytra, a warehouse-automation startup; Foxglove, a robotics software company; ForSight Robotics, which is developing a robotic system for eye surgery; and Blue Water Autonomy, a startup building autonomous naval ships.
"This is the very beginning," said Winterroth. "Anywhere that there is dangerous, dirty, and monotonous work being done in the physical world is ripe for robotic solutions."
Lux Capital
Farshchi's interest in robotics began with science fiction. He grew up watching Knight Rider and Star Trek, and said he was captivated by the idea that technology could give people superpowers. "A car that can drive itself felt like something that was achievable in my lifetime, even when I was a child," he said.
Decades later, as a partner at Lux Capital, Farshchi became an early investor in several companies that have defined modern robotics and autonomy, including self-driving-car startup Zoox in 2015 and warehouse-robotics startup Covariant in 2017. Both were later acquired by Amazon, in 2020 and 2024, respectively. He also backed Formic, a robotics-as-a-service startup that helps manufacturers deploy automation without paying for the equipment upfront.
Lux's broader robotics portfolio also includes Applied Intuition and Physical Intelligence.
"The zeitgeist today around robotics very much reminds me of the zeitgeist around self-driving cars back in 2012," Farshchi said.
Sequoia
Maguire's interest in hardware began at age 7, when his cousin showed up to Thanksgiving with a car full of computer parts and helped him build his first computer. He became obsessed with physics, earned a Ph.D. in quantum gravity at Caltech, and went on to found a space-launch company in 2009.
A few years ago, Maguire wrote an internal Sequoia memo he called the "Hardware Manifesto." His argument was that major software shifts depend on breakthroughs in hardware.
"You can't have the App Store without the iPhone. You can't have the iPhone without 20 years of work at Qualcomm and Broadcom building out cellular infrastructure. You can't have deep learning without the GPU," he said.
Maguire, who was also an early investor in SpaceX, believes AI will need a new generation of hardware to reach its full potential. He sees three forces pushing that shift faster than he expected: labor shortages, reshoring, and AI advances that make robots better at handling messy real-world situations.
His robotics investments include Neros, which develops unmanned drone systems for battlefield defense, and Mach Industries, a defense startup building unmanned hardware systems. He is especially excited about AMP Sortation, which uses cameras and robotic arms to identify recyclable items and sort them faster than people can.
Erin Beach
Before becoming an investor, Strohband helped build one of the most important early autonomous vehicles. He was the lead engineer and project lead for Stanford's autonomous car Stanley, which won the 2005 DARPA Grand Challenge and is now featured at the Smithsonian.
Participating in the challenge was "mildly terrifying," Strohband said, but it also showed him why autonomy was starting to become possible. Sensor prices were falling, and advances in probabilistic robotics meant machines could begin making decisions under uncertainty rather than relying solely on prewritten rules.
Strohband joined Khosla Ventures in 2012, and his first investment was Blue River Technology, an agriculture robotics startup that used computer vision to distinguish crops from weeds. John Deere acquired the company for $305 million in 2017.
He now thinks about robotics investments in two buckets: companies building entirely new technical capabilities, and companies using modern robotics to solve specific customer problems. His recent bets include Waabi, an autonomous-trucking startup; AIM Intelligent Machines, which is building autonomy software for heavy equipment; and General Intuition, a research lab using games and simulation to train AI agents for real-world tasks. Strohband and Khosla Ventures founder Vinod Khosla co-led the firm's investment in General Intuition.
Khosla has also been spending more time on robotics, with investments in Physical Intelligence, Field AI, and Genesis AI, which is building AI models and hardware for robots.
"Almost everybody in the 2030s will have a humanoid robot at home," Khosla said last year.
Founders Fund
Stephens took an unusual path to venture capital. After studying at Georgetown's School of Foreign Service and working for a US intelligence agency, he joined Palantir as an early employee, helping grow its defense business.
Peter Thiel recruited Stephens to Founders Fund in 2014. At a firm retreat, Stephens met Palmer Luckey, and the two developed the idea behind Anduril: bringing a Silicon Valley approach to advanced autonomous systems for defense. Anduril is now valued at $61 billion, and the company's president has said it is a "heavy participant" in the Middle East conflict.
Stephens gravitates toward investing in companies that combine software and hardware to solve real-world problems. His robotics investments include Hadrian, which automates precision manufacturing for aerospace and defense; Gecko Robotics, which uses robots to inspect critical infrastructure; and Physical Intelligence, which is developing general-purpose AI models for robots.
He believes that the biggest robotics companies will not be built on hardware alone, but by combining AI, software, and manufacturing know-how into systems that work reliably in the real world.
from Business Insider https://ift.tt/Q5mdrbq
