Jeff Cardenas shoots your MacBook. The Apptronik co-founder and CEO has a slideshow he wants to show, looking back on the Austin startup’s seven-year history. In fact, it requires a bit of contextualization. Like many other robotics companies, the company was driven by government contracts in its early days.
First up was Valkyrie 2, the second iteration of NASA’s humanoid space robot. The fledgling company was one of the few tasked with helping bring that system to life. Their contribution to the puzzle were liquid-cooled robotic actuators developed at the University of Texas Human Centered Robots Laboratory, led by Apptronik co-founder and chief scientist Luis Sentis.
Next up was exoskeletons. United States Special Operations Command (USSOCOM), which was in the market for “iron man suits.”
“[The]the exoskeleton was cooled with liquid”, says Cárdenas. “We learned a lot doing that. The complexity of the system was too high. It was heavy. We remoteized all the actuators. And then we started to realize what the simplest version of a humanoid robot was: a mobile manipulator. We started to get approached by a lot of logistics people, who didn’t want to pay for the manufacturing of weapons. They were too precise for what they needed. What they wanted was an affordable robotic logistics arm.”
Industrial weapons have been by far the spearhead, having been around for roughly 60 years. In that time, manufacturers like Fanuc and Kuka have engineered pinpoint accuracy. That is something that is essential for manufacturing and excessive for most logistics work. Installing electrical systems in cars, for example, requires much more precision than moving a box from point a to point b.
“What a lot of people are doing with humanoid robots is basically trying to build them the same way we’ve built industrial robotic arms for a long time,” says Cárdenas. “One of the key insights for Apptronik is that the way we need to build these robots, when you have a 30-degree-of-freedom system, is fundamentally different. The things we need are different. We need them to be safe around humans, we need them to be very resilient to the environment they’re in. We need them to be highly energy efficient. It’s a new set of constraints that we’re trying to optimize. Taking the same architectures from all the arms you see out there [at Automate] and extrapolating that doesn’t make sense. This is a fundamentally different architecture where we have about a third fewer components per actuator, it takes about a third the time to assemble.”
For a number of reasons, logistics is a logical place for a company like Apptronik to land. Not everyone wants to be sustained by government contracts forever. In recent years, logistics/fulfillment has become the most popular category in the robotics space. Like many other companies that have moved from research to the world of commercial products, the company had to determine if the right market existed for the technology it was creating.
“The goal was to get to the humanoid,” says Cárdenas. “The humanoid is a kind of holy grail. Probably the only thing that was consistent when we started was, ‘Don’t do humanoids. They are too complicated. ”
The true value of humanoid robots in the workplace remains an open question. But at least Apptronik isn’t the only one asking. Tesla’s much-publicized Optimus ad shook something loose. Suddenly, companies that had been operating in stealth mode felt compelled to announce their own intentions. Startups like 1X and Figure have discussed their progress to varying degrees. Sanctuary AI, which has partnered with Apptronik’s hardware, has already begun testing systems.
Apptronik, for its part, has so far shown two halves of a robot. There’s Astra, the humanoid robot upper body, which can be assembled into an autonomous mobile robot (AMR). At the other extreme is Draco, who is literally all legs. The company refers to it as its “first biped,” which is true, but that’s all there is to it.
Part of our impromptu slideshow are videos of the skinny legs walking around Apptronik Labs. He hasn’t matched Cassie’s speeds from the looks of her, but looking at him just with the naked eye, the pace seems faster than what Tesla showed off in his recent Optimus videos. What becomes clear after looking at a handful of these seemingly disparate projects is that Apptronik has been building its own complete humanoid robot piece by piece from the ground up.
Cárdenas says that the company was launched for a large part of its existence, until it reached about 40-50 people. It will explore a Series A this year, after the official presentation of its complete humanoid system this summer. “We have all these building blocks,” he adds. “A lot of it has been iterating and trying new ideas. The advantage of bootstrapping is that we have believed in it for a long time. We’ve been at this now for about a decade as a team, since Valkyrie.”
At the moment, the company is working on walking and developing the core functionality of the robot before launch. It is understandable that you want to prove that the product actually works as planned before you show it to the world. It’s a markedly different approach than what Tesla has been doing with Optimus, and if all goes to plan, it will propel the company to its next big raise.
Cardenas shows me images, both renders and photos, of Apollo, the system he plans to debut this summer. I can’t share them here, but I can tell you that the design contrasts with the kind of convergent evolution I’ve described, which found Tesla, Figure, and the OpenAI-backed 1X all rendered with a shared design language. Apollo looks, in a word, friendlier than any of these systems and the NASA Valkyrie robot that came before it.
It shares many more design qualities with Astra. In fact, you could even go so far as to describe it as having a cartoonish aesthetic, with an old-school iMac-shaped head and a combination of button and screen eyes making up the face. While it’s true that most people won’t interact with these systems, which are designed to work in places like warehouses and manufacturing plants, you don’t need to embrace the sinister for the sake of looking cool.
In some ways, the general purpose part is more difficult than the humanoid part. That’s not to say that building a fully mobile, articulated bipedal robot is far from easy, but there’s a wide gulf between special and general purpose. The precise definition of the latter is a conversation for another day, but for many, the label describes a system that is fully adaptable on the fly. For some, that means something like an API and app store for outside developers to build skills, but the systems still have to adapt to their environment. Ideally, it’s a machine that can do any task a human can do.
Too often, people don’t recognize the vast middle ground that is multi-purpose systems. At the moment, this is a much more pragmatic place to operate. Tesla’s notion of a robot that can work in the factory all day, do the grocery shopping, and come home and cook dinner feeds existing fanciful expectations that have been fueled by decades of science fiction.
“To get it to do multiple things,” says Cárdenas, “it’s still early days, but there are enough applications where if we can do simple things like move a box from point a to point b, there are tens of thousands of units.” value of claims for those applications.”
Like all work in space, these conversations require the caveat that we are still in the early stages. Agility has arguably gone the furthest in terms of testing the effectiveness of a humanoid (or at least bipedal) robot in a warehouse environment. But even they have a long way to go.
Still, the next few years will offer a fascinating glimpse of where these culminations of decades of research are headed.