Waymo’s Vincent Vanhoucken joins us on the podcast this week. It’s a great conversation for a few reasons. First, Waymo seems to be everywhere these days. Last time I was in SF, I found myself in various parts of the city, surrounded almost entirely by the company’s distinctive white Jaguars, and it honestly felt like there had been some tear in the reality matrix. Thankfully, they’ve since opened up highway trips, allowing the self-driving cars a chance to spread out a bit across the Bay.
Vanhoucken also has a long and interesting history with Alphabet AI initiatives, including a stint running Google Robotics. The more people I talk to, the clearer it becomes that Google has helped form this current moment in AI and robotics in interesting and unexpected ways.
Remember when the company bought a bunch of remarkably distinct robotics startups back in 2013? While it’s safe to say the master plan didn’t play out the way Google had hoped, the massive industry shakeup has had repercussions that are very much being felt to this day.
Similar points have been made about autonomous vehicles. A lot of people banked on things coming together a lot faster, and some very promising firms didn’t go the distance. But so much of the work done in autonomy, mapping, sensors, etc., has profoundly impacted robotics and AI. The more bearish among us suspect things will play out in similar fashion for humanoids, but that’s a conversation for a different column.
Of course, Waymo has probably fared the best of the bunch. Honestly, the fact that we’re even talking about the company in the present tense puts them well ahead of the pack.
“Everything is a journey,” Zack Jackowski tells me. “You’ve got to do it step by step.”
Atlas’ general manager could be forgiven for taking his own victory lap a week after the humanoid strode into the spotlight amid a grand reveal. If CES 2026 was, indeed, the year of the robot, the 6.2-foot, productized version of Boston Dynamics’ humanoid stood out among the rest.
But if a decade of speaking to roboticists facing down the infinite complexities of converting research into scalable production has taught me one thing, it’s that both Atlas and its creators have only just embarked on their thousand-mile journey.
“We're not going to start with the hardest task in the car plant,” Jackowski adds, describing the in-house pilots for the company’s third commercial product. “We have a tremendous amount to learn along the way.”
Those precious first baby steps were, themselves, more than a decade in the making. The concept of productization was seemingly nowhere on Boston Dynamics’ radar when the MIT spinout took an early PETMAN prototype out for a brisk walk for cameras on an industrial treadmill in 2014.
Even by the non-existent standards of the day, that bipedal robot wasn’t much to look at. Indeed, the mind contorts to reconcile a beefy rectangular torso mounted atop a skinny pair of hydraulic legs with any shared notion of what constitutes a “humanoid.” The armless, headless system was supported by a series of cables snaked down from a gantry system.
“We've seen more data in physical retail stores than probably anyone on the planet,” Brad Bogolea says. It’s tough to argue from the standpoint of sheer numbers. While Simbe’s CEO won’t reveal the precise number of Tally robots currently cruising up and down grocery store aisles, he suggests I extrapolate from the following datapoints:
More than 1,000 cities across 10 countries
Cities can contain multiple locations with Tally systems
While prominent hardware startups have grappled with the question of amassing a large enough tranche of data to sufficiently train physical AI, Simbe’s robots have been counting whistling while they work, counting bananas, and spotting missing soup cans. I’m racking my brain to identify a robot I personally interact with more often in the real world, audibly greeting Tally during my weekly trip to one of two Poughkeepsie ShopRites.
Last November marked a decade since the Bay Area company announced the first version of the robot via press release. The pitch lays things out in simple terms: misplaces, out of stock items, and missing shelves contribute to an estimated annual loss of $450 billion for retailers globally. Inventory/story walks are a necessary but tedious and resource-consuming solution. How about a robot that can autonomously cruise up and down the aisles, constantly updating a map of store shelves?
Seems like a straightforward proposition from the vantage of a robotics newsletter in 2026. Jump back a decade, before terms like “digital twin” and “embodied AI” started losing all meaning, and you’ll remember that there are no sure bets in this business. Bogolea notes that robotics funding was significantly more difficult to come by as Simbe looked to launch its first product
“You and I know the old days, and if we back up 10 years, very few venture capitalists were putting dollars into areas like robotics,” he says. “We were fortunate to have research institutions like Willow Garage and other things way back.”
The long-standing existential debate between a future full of single- or general-purpose robots spilled over into the embodied AI side some time ago. There are two distinct schools of thought on the matter — the first holding that different robot form factors will require bespoke models. The second – for which Skild.AI is a leading proponent — envisions a model that can adapt and grow to meet different robot requirements, be they industrial arm, humanoid, quadruped, or, say, Furby. Put more succinctly in the startup’s press material, “If there is a machine that can move, the omni-bodied Skild Brain will eventually be able to operate it.”
The Skild Brain isn’t there yet, but its creators just raised another $1.4 billion to speed up that process. On a call this week, cofounder and president, Abhinav Gupta, tells me that doesn’t mean the startup is waiting in the lab until its Swiss Army AI is fully realized. A version of Skild Brain has been deployed on robots currently operating in the real world. One Fortune 500 has used an early version on an existing system, and Gupta notes that a robot has operated in La Guardia airport with help from the company’s models — something to look out for when I fly out to the A3 Forum in a few days. Something else worth looking forward to: the full break down of my conversation with Gupta, coming up in next week’s newsletter.
This generation and its incessant job hopping. I am, of course, referring specifically to the latest generation of humanoid robots, which seem to be hunting down a new gig every few months or so. What happened to the career-long loyalty of industrial robot arms? Humanoids are, of course, very much in the figure-it-out phase, as the corporate world looks on with great interest and everyone tries to figure out exactly what scaling looks like. London-based Humanoid, which prides itself on speedy rollouts, had a pair of announcements this week. Just this morning, Siemens revealed that it has completed a two-week on-site pilot d with a Humanoid ‘bot.
The German automation giant put the robot to work transporting totes and placing them onto a conveyor at a rate of 60 per hour. A promotional video tied to the news reveals the Alpha Wheeled robot at work, utilizing an AMR base, rather than a pair of legs (no doubt contributing to the aforementioned quick rollout). On a call this morning, Humanoid Chief Strategy Officer, Alina Kolpakova told me that the wheeled version of the robot will be the company's main industrial focus. " [W]e're not rushing into bipedal and we are not seeing bipedal applications in an industrial setting," she noted. "We have very unique exceptional cases, for example the shipping industry. You can have a bipedal robot operating on ships, but compared to the material moving industry and manufacturing logistics, it's very small. So we're not focusing on these like rare cases."
Rolling out an AMR version was a much faster (and relatively safer) path to the market. Including Siemens, Humanoid has completed six pilots "of similar nature" (material/tote handling). Another German firm, Schaeffler, plans to deploy the robots on a wider scale. On Tuesday, it announced its intentions to put “hundreds” of the robots to work in its facilities. As for Humanoid's bipedal robots, they will be focused on future applications in the home. Pilots for that will likely start in a couple of years. Humanoid is also hoping to show you a "general purpose robot for industrial applications" in roughly two years. More from that conversation next week.
Seems we’re in the middle of Germany’s Humanoid Robot Week. Amid the above pair of Humanoid announcements, actual German robotics company, Neura, is teaming up with German engineering conglomerate, Bosch. Neura has been talking up humanoids for a while now. CEO David Reger, the spikier-haired gentleman pictured above, joined my panel on the topic a couple of Automates ago. While this deal does incorporate these systems, it’s different than the pair from the previous story. The companies will be using “advanced sensor suits” to collect data from Bosch facilities. In essence, this is largely an exercise in grabbing some of that precious real-world data for robot training. “As part of the collaboration,” the companies note in a release, “Bosch and Neura will co-develop AI-based core and functional software, as well as intuitive user interfaces.”
Vincent Vanhoucke (Waymo) - When it comes to AI at Alphabet, Vincent Vanhoucke has seen it all, from research scientist at Google Brain to head of Google Robotics to his current role at Waymo.
Rohit Sharma (True Ventures)- The investor discusses how the firm determines early stage successes and charts his family’s journey from India farm to Palo Alto boardroom.
Best Robot Videos of 2025 with TRI - We break free from the studio (my guest bedroom) to discuss the year's best videos with TRI's Erin McColl.
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.
