It’s never a good time to say "goodbye," but thank you to everyone in my life for allowing me to drop everything for a week and a half to attempt as much. For those interested in getting a sense of who my father was in one shamefully brief anecdote, he asked me at one point last week whether his death was going to “mess with [my] work schedule.”
In a word, “yes,” but that’s the nice thing about having folks like Becca and Carole around. Robot safety panels are important, don’t get me wrong, but some things take precedence.
Here's to the man who built pinewood derby cars with me, fixed watches for a living, collected power tools, and flew remote control planes in his free time — and I still managed to become a creative writing major at some hippy liberal arts school. But hey, I got around to tech in my own backward way. He also instilled in me the kind of work ethic that has me back at this conference and here with another newsletter. There's no "Automated" without "Tom" (it's right there in the middle of the name).
Thanks to everyone else who reached out with condolences, and while, yes, it’s been weird to immediately jump back into our biggest event of the year, I’m grateful to be back around people talking about robots.
In the meantime, if you have the means and the desire to contribute to Parkinson’s research, this is a good place to do so. You don't need to have experienced it personally to know what an ugly thing it is. That said, seeing the mental and emotional take diseases like these have makes you want to use any and all available platforms.
This edition hits inboxes as we're wrapping up the final day of Automate 2026. The show just keeps getting bigger, and I promise we'll have more leftover for next week, too. Speaking of next week, come watch me host a stage at the first-ever Davos Tech Summit.
“What’s worse than a moonshot?” Nicolaus Radford rhetorically posits. “It was much more, it was much worse than that.” NASA’s Robonaut program turned 30 this year. Officially kicking off in 1996, engineering work began in earnest on the humanoid astronaut the following year.
By the turn of the millennium, the team had produced the first version of the robot torso. A more “portable” Robonaut R1B arrived in 2002. None managed to slip the surly bonds of Earth. Nor, for that matter, did Centaur, which mounted the half-humanoid atop a four-wheel mobile for some theoretical off-planet off-roading (though it did spend some time in Arizona).
Robonaut 2, famously, had a bit more luck when it arrived in 2011. The system marked enough of an advancement over its predecessor to earn it a place aboard the International Space Station (ISS). R5, “Valkyrie,” arrived two short years later, designed with more terrestrial ambitions.
The bipedal humanoid was developed by the Johnson Space Center (JSC) Engineering Directorate to compete in that year’s DARPA Robotics Challenge Trials. Ultimately, however, a networking issue bested the robot, which failed to its marks. If that’s not a relatable human challenge, I’m not sure what is.
Radford worked on these and other robotics project during his 14 years at home of the original moonshot. He parted ways with the space agency in 2014 to form ocean autonomy firm Nauticus Robotics. In 2018 he did the same with components manufacturer, Jacobi Motors. It would be another six years before he was ready to jump back into bipedal humanoids with both feet. It no doubt helped to have 20-year Institute for Human and Machine Cognition vet and recent Figure CTO, Jerry, on board.
“There were a lot of pieces,” Radford says of Persona’s 2024 founding. “From Jerry and my point of view, we’ve known each other twenty years and we've always been kindred spirits when it came to the challenges of designing humanoid robotics. We've had a lot of mutual interests and actuation, and what are the best actuators for walking, and the torque speed envelopes, and the dynamics of locomotion.”
Here in robotics, there are, thankfully, institutes and organizations whose work isn’t wholly driven by the profit motive. These are places where a simple technological breakthrough is reward enough, and where studies entail more than just customer demographics. These bodies have played an outsized role in advancing the state of the art. These are circles where phrases like “ROI” and “go-to-market” are better avoided in mixed company.
At times, the world of venture capital feels similarly disinterested with turning a profit — at least not in any sort of hurried way. This is often true in deep tech, where technology goals make for long delays and long shots. Some problems are big and take time, requiring trucks full of money, long runways, and a whole lot of patience.
Self-driving offers a good example of what happens when longshots start to pay off, as they finally appear to be for Waymo in particular. The landscape in Google X that kicked off the Self-Driving Car Project 17 years back is now wholly unrecognizable. From countless startups to big names like Argo, Mobileye, and Google’s semi-top-secret car project, very few have survived long enough to see an entirely new round of companies emerge.
For many, the self-driving car offers a helpful roadmap, as physical AI and humanoid robotics firms break fundraising records left and right. Each round arrives amid huge promises, with every update couched as a major breakthrough. I, however, believe we are in the midst of a bubble, as many expect. What happens when the impressive demos don’t translate to actual dollars?
“Companies raise based on some sort of a financial forecast,” Cybernetix Ventures’ Fady Saad tells me on a recent call. “The financial forecasts basically predict in a year or two we will start generating revenue, right? So what happens is after a year or two, investors look at the forecast and realize that, revenues are not coming. What are we going to do? And this, this is when things start to slow down.”
Locus grew out of a major logistics problem. Quiet Logistics had been using Kiva robots before Amazon acquired Kiva and took the product off the market. Instead of returning to a manual operation, the team started building its own robotics solution inside the warehouse.
Aya Durbin (Boston Dynamics) - After helping deploy industrial robots at 6 River System, Aya Durbin is looking to do the same with Boston Dynamics' Atlas Humanoid.
Andrew Barry (Generalist AI) - Generalist AI's cofounder and CTO, Andrew Barry, discusses physical AI's difficulties and jaw-dropping breakthroughs.
Daniel Rausch (Amazon) - What does Alexa look like in the age of large language models? Amazon VP Daniel Rausch discusses how the smart assistant is continually evolving.
The SPAC is back — now in humanoid form. The special-purpose acquisition company was all the rage during the pandemic — as an alternative method for going public through a “blank check” merger with a company effectively created for that exact purpose. Agility Robotics this week announced plans to get publicly listed under the “AGLT” stock ticker via a reverse merger with Churchill Capital Corp X, which formed in February of this year. The deal would value the Digit maker at $2.5 billion pre-money equity, while creating $620 million of expected gross transaction proceeds, per Agility. Among other things, the merger is a way to raise a lot of money fast, for a few movers in a now-popular category looking to expand their influence on the physical AI side of the equation.
Here’s CEO, Peggy Johnson, “Humanoid robots are a critical driver of American technology leadership and the future of global industry. With category-defining commercially deployed humanoid robots operating in real customer environments today, Agility is at the forefront of a new era where safety-first, AI-powered technology can reliably work alongside people to bridge labor shortages, increase productivity, and strengthen the resilience of our supply chains. We believe humanoids are at a meaningful inflection point in commercial adoption, and we are focused on meeting growing customer demand, expanding deployments, and advancing our roadmap across robotics, physical AI, safety systems, and enterprise software. As adoption accelerates, we believe Agility is positioned to address a market opportunity across manufacturing, distribution, and logistics environments in the United States that is estimated by management to be approximately $1 trillion.”
I have an in-person meeting set up with the company a few hours from now, so expect more on the subject next week.
We’ve devoted a good chunk of the podcast to discussing how the push for self-driving laid the foundation for present-day robotics. Halos is another excellent example of the phenomenon. Pulling from “18,600+ engineering years of autonomous vehicle safety development” (per NVIDIA), Halos for Robotics utilizes the chip giant’s robotaxi operating system as a foundation for warehouse robot safety. Per the above image, Agility (yep, them again) is the first humanoid partner to bake in the offering. The release announcing the platform notes that its Digit systems are deployed in GXO, Schaeffler, and Toyota Motor Manufacturing Canada, and appears to reconfirm the company is working with Amazon, though there’s been little in the way of news on that front since early pilots.
Here’s NVIDIA VP (and early Automated guest) Deepu Talla, “Physical AI is transforming how factories, warehouses, and logistics operations work, and robotics teams need a unified safety architecture to scale autonomous systems into these environments. With NVIDIA Halos for Robotics, developers and system builders can harness NVIDIA’s proven autonomous vehicle safety foundation to develop safer robots faster and bring them into industrial operations alongside workers with greater confidence.”
Two things to know right off the bat. First: Kinisi Robotics was founded by Brennand Pierce, the former chief robotics officer at Bear. Second: The former used the latter’s mobile platform for its humanoid system. As Pierce has told me a few times over the past year, no reason to reinvent the wheel(s) here. Bear is best known for its restaurant robots, which are designed to augment wait staff by carrying dishes to and from tables. Kinisi makes a wheeled humanoid for factories and warehouse settings (though aside from mastering the physics, I don’t see why it couldn’t also bring a pair of arms to the front of a restaurant). Another fun fact: In addition to appearing on a panel at this year’s Humanoid Robot Forum, Kinisi also won top prize at Automate 2025’s startup competition. The company appeared at this year’s Modex as part of Bear’s booth, as well, which maybe should have been a giveaway for this week’s news, but I just assumed the founder knew a guy who knew a guy.
I’m not entirely sure how this happened, but it turned into venture capital week at Automated HQ. Call it “Series A Summer.” In addition to my conversation/upcoming podcast episode with Cybernetix’s Fady Saad, Rebecca surveyed some of the industry’s top investors to get a better sense of where their venture dollars are going. She spoke to folks from SOSV, Eclipse, and Prelude Ventures. Each one pointed to a larger trend that also turned into a takeaway from my chat with Saad: humanoids are cool and all, but their portfolios are focused elsewhere. Keep in mind, these are folks who have been focusing on the space for well before we inflated the current bubble. Here’s SOSV/HAX’s Duncan Turner, “My first go-to is, [robots that] do the things that humans can't do. That’s where there's big opportunities at the moment. That's sort of the area that seems to be a little bit forgotten.”
I’ve long had the sense that the robotics industry skipped a few steps when it went straight to humanoids. Missing links between AMRs and general-purpose bipeds have begun popping up over the past year. Locus Robotics’ recently announced Array system springs to mind, along with a spate of wheeled humanoid robots from companies like Reflex and Kinisi. The space between AMR and full-fledged bipedal humanoid has plenty of room to accommodate bidexterous manipulation, and suspect the lines between grasping AMRs and wheeled humanoids will only get fuzzier. Proxie 2 is a great example. As the above image shows, Cobot’s latest features a modular two-arm “option” (i.e., upgrade), allowing the system to “cover a wide array of human-scale, two-handed tasks including hospital resupply, warehouse kitting, life sciences lab operations, and manufacturing line tending.”
"For decades, deploying robots has meant choosing between mobility and dexterity, and always required custom software integration," says CEO Brad Porter. "Our second-generation Proxie brings all of that together in one platform we designed end-to-end. It moves, manipulates, and orchestrates its own work. The robots identify what needs to be done, announce what they're going to do, and then they just do it." The second-gen Proxie can move payloads up to 1,500 lbs and vertically lift up to to 200 lbs. Customers can rent the system at $5,000 a month.
A little over a year ago, collegiate food delivery service Kiwi announced it was going in big with the domain game by purchasing Robot.com. As the U.C. Berkeley spinoff noted at the time, the news was about more than just an expensive URL — it was a rebrand designed to project the breadth of its future plans, as it grew into additional form factors. This week at Automate, the company revealed its long-threatened humanoid, R-Noid. The system sports a head, two arms, and a wheeled base, filling out a line that already includes R-kiwi, R-kiwi+, and R-cargo.
Robot.com is casting a wide net with this one, including — and beyond — the warehouse, with its sights also set on hospitality, restaurants, and the home. "The future of work isn't fewer people. It's people freed from the parts of the job that grind them down, doing more of what they're good at," says cofounder and CEO, Felipe Chavez Cortes. "We build the robots that make that trade real, taking the repetitive physical work off your team so they can focus on craft, care, and the customer."
The unintended startup funding trend continues this week, courtesy of a piece on Crunchbase from my friend and former colleague, Mary Anne Azevedo. The data notes that 2025’s funding came in at $15 billion for the robotics sector, besting the former peak of $14.1 billion in 2021, when pandemic-related shutdowns were fueling a massive automation push. But that’s nothing compared to the $18.8 billion raised through the first half of 2026, as humanoids and physical AI firms are being minted unicorns, left and right.
More Automate hotness from Intrinsic this week with the release of a new modular work cell system. The reference design is created to comport with complex automated assembly tasks, and is targeted at a wide range of different-sized customers, up to and including manufacturing giant, Foxconn. This week at Automate, the Google X spinout is showcasing the system with a Fanuc industrial arm. Says Intrinsic CEO, Wendy Tan White, “The intersection of AI with modular production is going to unlock new value and operating models across the industry. When both physical workcells and software-defined robotics enable more flexibility, it allows manufacturers of all sizes to build in new ways, and fundamentally shifts the economics of production.”
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.
.png?upscale=true&width=1120&upscale=true&name=Rick%20Faulk%20Thumbnail%20(1).png)
.png?upscale=true&width=500&upscale=true&name=images%20(3).png)
