Well before I started my first robotics newsletter for TechCrunch, someone floated a similar idea. The project seemed impossible. I could only picture myself scraping together the dregs of industry “news” each week, hours before deadline.
These days, the opposite is true. The job of curation is more about what doesn’t make it into Automated every week. It’s about prioritizing news stories and stopping myself from adding another section every week so my editor Maureen doesn’t kill me.
Right now is an amazing time to be covering robotics and physical AI, and I’m glad to have a newsletter and podcast that allow me to bring that information to you as directly as possible. I’m also glad to have some backup.
You probably haven’t seen Rebecca Szkutak’s byline on here, on the other hand — but you will soon. I’m thrilled to report that we’ve brought her on board our growing news team. Rebecca and I worked together at TechCrunch for about four years, and she took the lion’s share of the site’s robotics coverage after I headed out to launch this thing.
She’ll be helping expand our coverage of startups and VC. As I said, it’s an amazing time to be covering robotics and physical AI, and much of that excitement is centered around new companies that are rethinking the industry. In fact, here’s a list of 10 of them that were just announced as finalists for this year’s Automation Startup Challenge (June 22 at Automate in Chicago).
Someone once told me, “Brian, you can’t do everything yourself.” While I’m still convinced that person was lying to me and I intend to prove them wrong until I pass out from exhaustion at my standing desk, it’s good to have backup from some great writers. I continue to be amazed at how quickly Automated has grown, and I’m excited for what’s next (barring passing out at my desk).
Like most memorable phrases of uncertain origin, “boil the ocean” has been variously attributed to Mark Twain, Will Rogers, and Lewis Carroll — three great minds that appear to have done some of their greatest work long after death. Regardless of who first suggested we curb our seafood consumption, it’s long been an idiomatic yield sign for ambitious founders.
Pragmatism has never really been Silicon Valley’s strong suit, of course. One thing about moonshots is you really only need one or two to pay off to be considered a kind of sage. The other thing about moonshots, of course, is that even the ones that pay off tend to take a long time to do so, all while sacrificing a ton of hypothetical ROI along the way.
Somewhat in keeping with its name, Humble Robotics sits at an appealing cross-section of the pragmatic and wildly ambitious. A veteran of Mission Motors, Otto, and Spark AI, founder and CEO Eyal Cohen has spent much of his career on the latter side of the equation. Most recently, he worked as the head of hardware engineering, system engineering, operations at self-driving truck darling, Waabi.
Humble’s offering doesn’t stray too far from his previous employer’s autonomous ambition, but the company’s initial aim is — quite literally — focused on a significantly shorter road.
“Over the years, I’ve seen how much freight moves relatively short distances,” Cohen recently told me. “All kinds of freight is moving one to five miles. It's very short links of warehouse-to-warehouse moves. Sometimes you'll see factories moving like parts from one of the buildings to another. But they're carried by trucks.”
Where most autonomous trucking companies have focused on the massive long-haul market, Humble's approach targets a smaller, but more grounded market. Trucks traveling these short distances often do so at 20 to 30 miles an hour, and manage to avoid highways altogether.
“Tesla has one of the most automated factories in the whole world,” Joshua Joseph told me recently, “and automating material workflow is [one of the final pieces] of the puzzle.”
The Tesla manufacturing engineering is leading the carmaker’s efforts to fill that final piece, transitioning factory floors from the more traditional AGV (autonomous guided vehicle) model to AMRs (autonomous mobile robots).
In a recent piece for Institute of Industrial and Systems Engineers, he positions material workflow as a kind of unseen productivity bottleneck, writing, “Every finished product carries a hidden footprint: the distance its components traveled before reaching assembly. That travel governs cycle time, space use, and safety exposure. Even in highly automated plants, 25% to 40% of production delays trace back to poor material delivery.”
Those numbers, in turn, are pulled from a report about the Toyota Production System (TPS), “a production system based on the philosophy of achieving the complete elimination of waste in pursuit of the most efficient methods.” The Japanese carmaker breaks this into two factors. The first is jidoka, which boils down to automation with humans-in-the-loop to help guard against edge cases and other long-tail concerns.
The second factor, just-in-time, deals more directly with material workflow and is “based on the concept of synchronizing production processes.” More specifically, it’s about ensuring that only the necessary amount of product is produced, and that customers receive orders in as timely a fashion as possible.
The Dean of Carnegie Mellon’s School of Computer Science breaks down the university's advances in autonomy and offers insight into what the future holds. Watch on YouTube>
Bren Pierce (Kinisi) - After finding sustained success with restaurant robotics firm, Bear, Pierce is taking on humanoids with Kinisi.
Amit Goel (NVIDIA) - Recorded live at GTC, NVIDIA's head of robotics and edge computing ecosystem discusses the company's physical AI strategy.
Ali Kashani (Serve Robotics) - Serve's CEO on the journey from DoorDash experiment to Uber spinoff to delivery powerhouse.
On June 22, a handful of automation startups will compete for bragging rights, a large novelty check, an interview on Automate Live, and 10,000 real U.S. dollars. The Automate Startup Challenge highlights the next generation of brains and brands transforming this industry. Ten finalists will pitch their companies on stage at Automate in Chicago, while a panel of industry experts whittles the list down to one winner. It’s easily one of the conference’s most electric elements, offering a peek into what’s next for automation, manufacturing, robotics, and AI.
Since taking the top prize in last year’s contest, Kinisi has been making its mark with its wheel-based KR1 system. Just last week, founder Brennand Pierce was the guest on the Automated podcast. The week prior, the startup was among a very small number of firms with a functioning humanoid at one of the industry’s major trade shows. Today we’re excited to unveil the list of finalists set for the 2026 Automation Startup Challenge. Get to know a bit more at the link below.
According to Skydio CEO, Adam Bry, the most significant thing about the company’s $110 million Series F is “how little [it’s] raising.” It’s similar to the approach I took the last time I asked for a humble $110,000 raise, though my pitch didn’t go over nearly as well. Thankfully for Skydio, the company has established itself as the U.S.-based alternative to drone giants like DJI — not a bad spot in this era of government bans. Unsurprisingly, the firm has found particular success in public sector fields, including defense, safety, and infrastructure here in the States. All told, it says it has shipped north of 60,000 drones to nearly 4,000 customers. The list includes 1,200 public safety agencies, 450 energy/utility companies, and various militarily aligned countries. This new round values Skydio at $4.4 billion, and finds the company announcing a planned $3.5 billion domestic investment over the next half-decade as it further expands manufacturing. It claims those numbers will create more than 2,000 jobs within Skydio, along with 3,000+ roles in the U.S. supply chain.
You get $110 million, and you get $110 million. This Series B comes amid an uptick in activity for German robotics companies. It finds Sereact rolling out version 2.0 of its Cortex VLA model, which was trained on north of one billion real-world picks. The company distinguishes its approach as “plan-and-try,” as opposed to “try-and-see,” which effectively finds the system gaming out various outcomes before engaging in an activity. “Every successful pick, every failure, every recovery is captured with synchronized observations, robot state, gripper force feedback, and outcome — then filtered, prioritized by novelty and uncertainty, and used to update the model,” the startup writes. “Updated policies pass automated regression checks and roll out to the fleet. The loop closes. Data compounds. Coverage of the long tail expands.” Like sands through the hourglass.
Pudu this week announced a “nearly” $150 million round that brings its full to-date funding to $300 million and places its value north of $1.5 billion. The Shenzhen-based firm produces robots for a broad range of fields, including reality, hospitality, healthcare, and more, shipping around 120,000 systems annually, per its reporting. Pudu timed the funding announcement to coincide with the opening of its new U.S. headquarters in Dallas. The company says some 15,000 of its robots have been deployed across the Americas thus far, with a focus on delivery, cleaning, and the industrial sector.
Construction is a massive, global industry. It’s a prime candidate for disruption by robotics and physical AI, but has thus far lagged behind categories like manufacturing and fulfillment. What is the space missing, exactly? Could it be a large, six-legged bug with a robot arm mounted on top? RTP Global’s betting big that change will look something like the Mantis robot, leading All3 Robotics’ $25 million seed. The robot sports a reconfigurable tool designed to install, inspect, fashion, and finish work on a job site. “We're not building a universal robot,” the London-based firm’s head of robotics, Napo Montano said, “we're building the robotic solution that construction needs right now. Mantis addresses one high-impact problem: how to make custom, high-quality buildings affordable and scalable through tightly integrated automation.”
Automation’s long-time promise of streamlining life has been hindered by the time-consuming process of hard-coding robots. Getting a bunch of systems to do the same or similar things isn’t as simple as flipping a switch — particularly when each system features a different embodiment. New work out of EPFL’s School of Engineering in Zurich points toward a new method for transferring skills across different systems and physical configurations in a way that could significantly streamline the deployment process. The new robotic control framework, deemed “Kinematic Intelligence,” converts human demonstrations into general movements that can be shared across systems.
“A lot of recent work tries to address this using larger AI models and more data,” Sthithpragya Gupta, one of the work’s senior researchers, told A3. “While powerful, these approaches can be data-intensive, computationally expensive, and sometimes unpredictable in safety-critical settings,” says Gupta. “We took a step back, and instead of adding more complexity, we wanted to see if we can better understand what fundamentally governs how robots move. This led us to focus on robot kinematics, where the structure of its constraints enabled reliable skill transfer.”
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.
.jpg?upscale=true&width=1120&upscale=true&name=2216x1244%20(1).jpg)
