Supercharged by AI, humanoid robotics is garnering lots of attention from tech titans like Jeff Bezos and Elon Musk, so we took a deep dive into the sector in this month’s cover story, Robots to the rescue.
Relative to the overall robotics market, humanoids are still in their infancy. But the technology is advancing quickly, thanks in part to increased venture investment. VCs invested nearly $1 billion in 11 deals in the sector as of May 31, the largest amount in seven years, according to Crunchbase.
And there are signals that VCs believe the market will bear fruit in the near term. Earlier this year, humanoid robot maker Figure AI raised $675 million at a $2.6 billion valuation from Parkway Venture Capital, partnering with Microsoft, Nvidia, Evolution VC Partners and Bezos Expeditions, among others.
Driving demand for humanoids is a growing number of unfillable jobs in manufacturing and warehouses. At least 600,000 of those jobs are currently vacant and experts say the number is just going to keep getting bigger. Goldman Sachs foresees a $38 billion market by 2035.
For this month’s cover story, we spoke to experts from Parkway Venture Capital, DCVC, Anzu Partners, TUM Venture Labs, Creative Ventures, Stanford University Graduate School of Business’s Robotics Center, Interwoven Ventures and others about the prospects for this technology.
While there is a lot of enthusiasm around humanoid robotics, many of the VCs I spoke with are skeptical of some of the start-ups being pitched to them. Fady Saad, co-founder and general partner at Cybernetix Ventures in Boston, sees a lack of clarity at several robotics start-ups about what problem they are trying to solve as a major challenge to commercialization.
Without knowing the particular market-driven use case you are trying to address, “You’ll be over-engineering the machine that you are building,” Saad told me. “Do you care about the dexterity? Do you care about the speed? Do you care about the payload? Do you care about navigation? You cannot answer this [commercialization] question unless you know what problem you’re trying to solve.”
For example, a lot of effort is going into development of tactile sensors that will make robotic hands more versatile. “I spoke with many of those companies, and honestly speaking, many of them are having a hard time commercializing,” Saad noted. “Because again, where are we selling this? Do we start with the grippers? Can we add those tactile sensors to the grippers so the gripping mechanism knows exactly what it’s gripping and how much force they should use? And now they’re switching to humanoids, but I think both are technologies looking for problems [to solve], so it wouldn’t go far.”
Mind you, Saad is no robotics detractor. He co-founded MassRobotics, a nonprofit that is dedicated to accelerating robotics technology, innovation and commercialization.
‘An astronaut and a humanoid’
One business case Saad is convinced makes a lot of sense is space exploration. “Now you have a lot of reasons to design a humanoid,” he said. “First, if it’s smart enough and you can send it instead of an astronaut, maybe instead of sending two astronauts, you’re sending an astronaut and a humanoid. And now you need it to be human-like because you don’t want to change the controls [or the interfaces] because it will be co-locating with a human being.”
To navigate and explore uncertain topographies on other planets, a bipedal robot would be very helpful. “Because now if it falls down, it would be able to get up,” unlike the wheeled robot that recently landed on its side on the moon and was rendered useless, Saad explained.
Even if the cost to build such a humanoid were as much as $500,000, Saad believes it would be far cheaper than training and sending a human into space. He thinks this may be why Bezos and Musk, both of whom have pursued commercial space projects, are either developing their own humanoid technology or investing in companies developing the tech.
Version 3 of Tesla’s Optimus robot – an iteration that is said to maximize utility and be geared more to external consumers than Optimus 2 – could come out as early as September. Meanwhile, Bezos Expeditions has made a big bet on Figure AI’s humanoid, which is now being tested in BMW’s manufacturing plant in Spartanburg, South Carolina, and will soon be tested in some Amazon worksites as well.
Both Musk and Bezos understand that we “cannot go to other planets unless we have humanoids,” Saad said.
Go deeper into this subject by reading our cover story here.