“The previous leg configuration was really optimized for very good walking and running,” Jonathan Hurst said. “The new configuration is optimized for lifting heavy things and getting up and down off the ground a lot.”
Hurst is Agility Robotics’ co-founder and chief robot officer, and he is describing the single most visible thing about Digit 5: the ostrich-style backward knees are gone.
Last issue, this thread was about humanoid margin disclosure — what these companies actually book against what they claim. This week it gets both a product generation and a listing vehicle. Digit 5 arrives as Agility moves toward a proposed SPAC merger with Churchill Capital Corp. XI, valuing the company at $2.5 billion.
A product generation and a listing vehicle arrive together: Digit 5's lifting redesign now faces the harder test of public-market numbers.
The leg redesign is not styling. The new legs run on proprietary cycloidal actuators and can repeatedly lift up to 50 lb. (22.7 kg), which Hurst says covers single-person lift tasks in OSHA-regulated facilities.
“The product requirements push us in the direction of lifting heavy things all the time,” he said. “Then, as we try to do the engineering to solve this new priority, we end up with something that looks a bit more humanoid.”
He is explicit that this is a byproduct, not a goal. “It’s really important that we are not trying to copy what a person looks like or what an animal looks like. This is where the engineering took us.”
The bird legs were never arbitrary either. Cassie, Agility’s pre-Digit biped, arrived at that configuration because the team was chasing ostrich performance: 40 mph, all day. In 2022 Cassie ran 100 meters in 24.73 seconds at Oregon State’s Whyte Track and Field Center and took a Guinness World Record for the fastest 100 meters by a bipedal robot.
That record is now a historical footnote to a lifting spec.
The safety architecture depends on the new legs too. Digit 5 has padding on its knees for kneeling, and it can squat into a stable position — sit down and cut its motors when a person approaches.
“The robot needs to be in a statically stable post and be able to turn off its motors before a person can touch it,” Hurst said.
That is the mechanical half of the claim that Digit 5 can work in close proximity to people without physical safety barriers. Agility says there are no defined zones: as people get closer, the robot evaluates risk, slows, and eventually stops.
This matters because the few humanoids being tested in industrial facilities today generally are not working near human workers. Agility calls Digit the first functionally safe humanoid, and frames it as the first step toward robots that work with humans rather than merely near them.
The less glamorous numbers may matter more. Digit 5 runs 90 minutes and fully charges in nine, a 10:1 run-to-charge ratio, up from Digit 4’s 2:1. That supports more than 20 hours of work in a 24-hour day.
The robot is also bigger: 284 lb. (129 kg), 5 ft. 11 in. (1.81 m) tall, reaching 7.2 ft. (2.2 m) versus Digit 4’s 5.5 ft.
And it takes swappable grippers, which turns a tote handler into a platform that can be re-tooled for depalletizing, machine tending, kitting, inspection, and palletizing. Hurst is blunt about why: “One thing we learned is that general-purpose hands don’t exist. There are none on the market.”
The design process behind all of this is the part Agility keeps emphasizing. Earlier Digits were proof-of-concept machines. Working with Amazon on putwall bin removal taught the team that doing the task was the easy half.
What followed were FCC requirements, UL certifications, ISO certifications, aisle widths, where chargers can be bolted down, and what the workflow actually looks like. Hurst says those constraints turn into hundreds of engineering requirements across 10 use cases on the roadmap.
“That is the foundation of the design for Digit 5,” he said. “That’s not something we’ve ever done before, organizationally.”
So the pitch is: real facility constraints, not a demo. The financials are where that pitch gets tested.
Agility says it has more than $300 million in multi-year orders across manufacturing, warehousing, and logistics. Recent SEC filings put 2025 net sales at $1.8 million. The SPAC values the company at $2.5 billion.
That gap — $1.8 million booked against a $2.5 billion valuation — is exactly the kind of thing last issue’s margin thread was built to hold up to the light. Multi-year orders are not revenue. Revenue is not margin. And the company is asking public markets to underwrite a buildout that starts with early access to Digit 5 in the first half of 2027 and general availability by the end of 2027.
There is also a deployment-experience number worth sitting with: 65,000 hours of Digit field experience. At 20 hours a day, that is 7,300 hours per robot per year, so the entire deployed history comes to roughly nine robot-years. That is the accumulated evidence base behind a redesign that claims to cover single-person lifts across an OSHA-regulated facility.
What the public record does not contain is throughput. A reply to Agility’s own post asked the obvious operational question: cases per hour, and how often a human has to take over. No published task-hours figure accompanied Digit 5’s launch.
The rest of the go-to-market is geographic. Digit 5 is Agility’s first commercial deployment push outside North America, starting with the EU and U.K. The company expects a CE mark and additional regulatory marks for the EU, Great Britain, and Northern Ireland.
Digit 4, by contrast, was deployed with Toyota Motor Manufacturing Canada, GXO Logistics, Schaeffler, and Amazon. Digit 3 could stack boxes up to 40 lb. (18 kg), and Ford bought the first two robots off the production line in 2019 to explore van-to-doorstep delivery.
Agility now has more than 400 employees and a factory capable of building up to 10,000 robots annually. The open roles tell you where the remaining risk sits: a staff firmware engineer for motor controllers, limb stack boards, and sensor integration through real-time EtherCAT, and a senior PM for perception and world understanding, moving capabilities from research to prototype to production on a deployed humanoid.
Hurst’s framing is that this is a step on a long road. “Being the first commercially deployed humanoid robot, and being the first functionally safe humanoid robot, those are important steps on getting to the home.”
The legs are the headline. The ratio to watch is the one nobody has published yet: lifts per shift, and interventions per thousand.