Prior to this week, the only Houston I spent any time around is the street separating NoHo from SoHo. I’ll have to give you the full break down on the differences between the two if I ever make it through the TSA lines alive. One thing I can tell you for certain: neither reacts particularly well if you pronounce it like the other.
I’m here in the birthplace of Beyonce and Billy Gibbons to emcee and moderate the occasional panel at my first International Robot Safety Conference (IRSC). As evidenced by the above bingo card, my brain is awash with a seemingly random assortment of letters and numbers. Again, more thoughts on that later, after I'm able to hit the big red button on this week (sleeping at home, in my bed).
Exploring the nuances of robotics’ safety has been one of the more interesting avenues of A3-specific work since starting at the association back in March. It goes deep, and I happened to step in at a time when that wing of the org was generating a lot of public-facing news, including what we’ll – for brevity's sake – call the “humanoid safety standard," as well as the big update to R15.06 and its eternal battle with “cobot.”
More soon. In the meantime, there’s always this recent Automated episode with Roberta Nelson Shea from a few weeks back, which will bring you up to speed with much of the recent goings on in the world of robot safety.
I want to talk about Moravec’s Paradox, so I put the question to LinkedIn. A handful of responses suggest a handful of experts, including recent Automated guest, Rodney Brooks, and the eponymous Hans Moravec himself (thanks Helen G.). The Austrian computer scientist remains on my guest wish list, at the moment. For the sake of keeping to my weekly newsletter cadence, I instead reached out to an expert in the field I’ve spoken with frequently over the past several years (including, spoiler: next week's Automated).
I emailed Berkeley professor and Ambi Robotics chief scientist Ken Goldberg a note with the pertinent paradox serving as the subject line. “Hi Brian,” Goldberg responds almost immediately, “that’s funny because I was just talking about him.” Before replying, I take a moment to consider the number of people on Earth who could have genuinely responded to my initial message in this manner. I imagine it’s a fairly small club.
As for why I, specifically, was thinking about Mr. Moravec and his wonderful paradox, the short answer is humanoid robots. Humanoid robots, expectations, and the disconnect between what these unquestionably technologically impressive machines can and cannot do, right now.
Even Goliath, one imagines, began life as a David. And as for David, well, he went on to become a giant in his own right, before all was said and done. Herein lies the tension at the heart of the Roomba story. Around the turn of this past century, iRobot played the David role to Electrolux’s Goliath, as the Massachusetts startup and Swedish appliance giant suddenly found themselves swimming in the same waters.
We know how that chapter ends, of course. YouTube wasn’t awash with videos of baffled cats riding tricked-out Electrolux Trilobites during the 2010s. After several failed attempts to put itself on the map, iRobot won that battle, full stop. A second Trilobite was released three years later, before the product went the way of its arthropodic namesake.
Joseph Jones’s Dancing With Roomba recounts the moment in consumer robotics history directly from one of its principal creators. It’s a rare robotics book that’s more memoir than technical manual, going so far as recounting a little bit of quasi-industrial espionage as a treat.
According to the most recent publicly available figures, Waymo currently operates 800 cars in San Francisco. When I was in the city two weeks back, there were moments I could swear they had them all circling the block I was on, like I was the target of some elaborate VC prank. For what it’s worth, as The SF Examiner noted in August, that number is more than 3x the last one the Alphabet-affiliated automotive company shared publicly. A few months earlier it added, “We’ve also incrementally grown our commercial fleet as we’ve welcomed more riders, with over 1,500 vehicles across San Francisco, Los Angeles, Phoenix, and Austin. Earlier this year, we received our final delivery from Jaguar, and through next year, we will build over 2,000 more fully autonomous I-PACE vehicles for our fleet.”
Detroit, Las Vegas, and San Diego get ready, because those cars are driving themselves your way — though Waymo has yet to offer a firm date. The company had previously referenced its desire to launch in a series of additional U.S. cities including Miami, Seattle, Boston, Denver, and New York, though each deployment comes with its own unique local and state regulatory red tape. The company is facing mounting government pushback in Boston, while San Francisco officials mourn the loss of a beloved bodega cat. RIP KitKat. May his memory be a blessing.
Friend and former colleague Mary Ann Azevedo pointed me to the $91 million raise you’ll find below — but I wanted to pull on a different thread before we get there. Specifically, I want to talk about the metrics that make up that above chart. The data paints a fascinating story of robot funding, now that we’re a bit over halfway through the 2020s. The info was mined from Crunchbase, and as the fine print notes, raises have to be at least $200,000 to be included in the "total rounds" metric. With just under two months to go, robotics venture has been funded to the tune of $10.4B, whizzing past the whole of 2024 ($7.5B), 2023 ($7B), 2022 ($8.7B), and 2020 ($5.7B). That’s a 36% increase in the last year alone (and again, it’s still early November).
Of the past five years, only 2021 ($13.8B) was more fruitful for robot startup coffers. An interesting wrinkle in all of this is that — in spite of total funds being up significantly — 2025’s 376 rounds (so far) lag behind each of the last five years, culminating with 408 in 2024. This means, you guessed it, that the average rounds are getting larger. Of course, anyone who’s been clocking embodied AI and humanoids over the past 10 months probably could have told you that. As Azevedo notes, however, humanoids aren't the only category pulling in huge rounds. They're joined by “surgical robotics, operating systems, and manufacturing automation." They’re a geographically diverse group as well, spanning the United States, Europe, and China.
For years, the question of what drones can actually do in real world terms largely revolved around imaging. Mind, that one category covers a lot of applications across a lot of industries, from inspecting power plants, to surveying wild life, to snapping photos for real estate listings. Attaching additional mechanisms to a system for the purpose of increasing functionality, while promising, complicates things. In a video, Infravision CEO Cameron Van Der Berg notes that the proposal to string power lines with big quadcopters was met with its share oh naysayers along the way. For now, the Austin-based firm is taking a victory lap to the tune of $91 million in fresh Series B funding. Infravision has already teamed up with a number of big utility companies, including California’s PG&E, for which it executed emergency response deployments.
We’ll be debating what qualifies as a humanoid until the worms are munching on our own corporeal forms. I can, however, tell you with at least six degrees of confidence that a pair of robot arms and hands does not a humanoid make. And — depending on your specific needs — that may be perfectly fine. In fact, ETH Zurich grad Mimic Robotics is betting that it will be more than enough for most industrial tasks. “Humanoids are exciting, but there aren’t many industrial scenarios where the full-body form factor truly adds value,” chief product officer, Stephan-Daniel Gravert, says in a release. “Our approach pairs AI-driven dexterous robotic hands with proven, off-the-shelf robot arms to deliver the same capabilities in a way that is much simpler, more reliable, and rapidly deployable.” Mimic just raised a $16 million Series B to speed up the development of its foundational model and human-style robot hands, along with extending those robot arms’ market reach.
A promising, but short-lived push into open-source humanoid robotics is shutting down. Bay Area-based K-Scale Labs will wind down operations, owing to an inability to secure sufficient funding. The Information was the first to report the news, by way of an email CEO Ben Bolte sent to the firm’s existing investors. The former Tesla and Facebook engineer noted that K-Scale was unable to secure the $10 to $15 million the startup had aimed for, and was instead winding down with around $400,000 left from a $4 million raise, per the report. Prior to informing shareholders and sharing the news via the company’s Discord server, Bolte writes in the letter that the team unsuccessfully explored a sale to both 1X and The Bot Co. The former, at least, apparently had little interest beyond hiring a handful of existing staff.
K-Scale had taken a decidedly different approach than a majority of the humanoid market, open sourcing code via Github, and offering up its bipedal K-Bot system at an $8,000 price point over the summer. In July, the Palo Alto firm announced that it had hit $1 million in orders. In the note, the CEO says the company will refund the unfulfilled pre-orders.
The Association for Advancing Automation (A3) is North America’s largest automation trade association representing more than 1,400 organizations involved in robotics, artificial intelligence, machine vision & imaging, motion control & motors, and related automation technologies.
%20-%20Edited%20-%20Edited.jpg?upscale=true&width=1120&upscale=true&name=unnamed%20(2)%20-%20Edited%20-%20Edited.jpg)
