Agility Robotics' DIGIT: The Humanoid Robot Revolutionizing Warehouse Work

by Marcin Frąckiewicz
in Uncategorized
on 31 August 2025

Key Facts

First humanoid in real warehouses: Agility Robotics’ DIGIT is a bipedal humanoid robot (about 5’9” tall, ~140 lbs) designed for logistics work reemanrobot.com. It’s the first commercially deployed humanoid robot operating in warehouses and factories, performing tasks like moving and stacking tote bins in supply chain facilities techcrunch.com techcrunch.com.
Human-like mobility and manipulation: DIGIT walks on two legs and has multi-jointed arms with advanced grasping. It can carry 35-pound loads (with next-gen models targeting 50 lbs capacity) reemanrobot.com, navigate human environments (e.g. steps, tight spaces) technologymagazine.com, and handle objects on shelves or conveyors. A suite of cameras, LiDAR, and sensors provides 360° vision and obstacle detection geekwire.com, enabling autonomous navigation without needing pre-mapped routes agilityrobotics.com.
Robots-as-a-Service pricing: DIGIT is available for commercial use now. Agility offers it via a Robots-as-a-Service (RaaS) subscription, roughly $30 per hour for a fully managed robot (including software and support), yielding customer ROI in under 2 years reemanrobot.com. Companies can also opt to purchase robot fleets outright, supplemented by software/service fees agilityrobotics.com.
Industry adoption has begun: Logistics leader GXO is deploying DIGIT in a Spanx distribution center to automate tote handling techcrunch.com. Amazon has piloted DIGIT for tote consolidation tasks in its warehouses geekwire.com. Agility reports that multiple customers in warehousing, retail, and third-party logistics are testing or using DIGIT to alleviate repetitive material-handling work techcrunch.com technologymagazine.com.
2025 upgrades boost endurance & safety: New enhancements (unveiled at ProMat 2025) double DIGIT’s battery life up to ~4 hours active operation agilityrobotics.com (up to 8 hours under light use reemanrobot.com) and enable autonomous charging docking agilityrobotics.com. DIGIT’s latest version also features improved safety systems – on-board emergency stop buttons, safety-rated PLC controls, and status displays – aiming for “cooperative” human-robot work compliance agilityrobotics.com technologymagazine.com. Robust new limbs and end-effectors expand its range of motion for more manipulation use cases agilityrobotics.com.
Scaling up production: Agility opened “RoboFab,” the world’s first dedicated humanoid robot factory, in Oregon and plans to manufacture 10,000+ DIGITs per year in the next few years reemanrobot.com. The company (led by CEO Peggy Johnson since 2024) raised major funding (including backing from Amazon) to ramp up R&D and meet growing demand geekwire.com geekwire.com.
Crowded humanoid race: DIGIT faces emerging competition from tech giants and startups. Tesla’s Optimus, Figure AI’s Figure 01/02, and others like Boston Dynamics’ Atlas (research platform) and Stretch (box-handling robot) are all vying to redefine automation. However, as of 2025 DIGIT is ahead in real-world deployment, being one of the few humanoids generating revenue on real tasks techcrunch.com. Analysts predict the humanoid robot sector will grow explosively – Goldman Sachs forecasts a $38 billion market by 2035 goldmansachs.com – as these robots mature from prototypes to practical workers.

Meet DIGIT: A Humanoid Built for Work

DIGIT is Agility Robotics’ flagship humanoid robot, purpose-built to tackle physical work in human environments. Standing about 1.7 m (5’9”) tall and weighing ~64 kg (140 lbs) reemanrobot.com, DIGIT has a human-like bipedal form factor. It walks upright on two legs, giving it the ability to traverse terrain and spaces designed for people – from ramps and doorways to narrow warehouse aisles technologymagazine.com. DIGIT’s arms serve both for balance and manipulation; they’re equipped with grippers enabling it to pick up and carry containers, boxes, and other objects up to about 35 pounds in weight reemanrobot.com. (Agility has indicated an upgraded model due in late 2024 will increase payload capacity to 50 lbs reemanrobot.com.) The robot’s torso contains its battery and computer “brain,” while its head is adorned with LED “eyes” that signal status and direction of movement (for example, blinking to show which way it’s about to turn) geekwire.com – a friendly touch to make human coworkers more comfortable.

Sensor & AI suite: To operate autonomously, DIGIT is loaded with sensors. Multiple stereo cameras and depth sensors give it a panoramic view to detect obstacles and locate objects. A scanning LiDAR unit mounted in its head lets DIGIT map its surroundings in 3D and navigate with centimeter-level precision geekwire.com. These onboard sensors feed into DIGIT’s AI software, which uses deep learning for vision and decision-making. Agility has integrated NVIDIA’s Jetson AI computing platform in DIGIT, leveraging powerful GPUs for real-time perception and control algorithms geekwire.com. In fact, Agility recently announced it will adopt NVIDIA’s next-gen Jetson Thor platform to push DIGIT’s capabilities further, enabling more complex on-robot AI models for whole-body control, manipulation, and scene understanding agilityrobotics.com agilityrobotics.com. Thanks to this “full-stack” approach – tightly coupling hardware sensors, actuators, and AI software – DIGIT can roam in new environments without prior mapping and adapt on the fly. It plans its footpaths and arm motions dynamically, avoiding sudden obstacles (like people or forklifts crossing its path) by stopping and then resuming once clear agilityrobotics.com.

Balance and locomotion: DIGIT’s distinctive leg design (with reverse knees reminiscent of an ostrich or dinosaur) stems from decades of research into bipedal locomotion geekwire.com. This design gives it a nimble, dynamic gait. It can recover from small stumbles, step over low obstacles, crouch to reach lower objects, and climb stairs if needed. While not built for acrobatics like Boston Dynamics’ Atlas, DIGIT is robust enough to handle the bumps and uncertainties of a warehouse: its control system automatically adjusts joint stiffness and position to keep balance if it collides with an object or gets nudged agilityrobotics.com. Such whole-body control is critical – DIGIT must maintain stability while carrying awkward loads or reaching out with its arms. The payoff is a machine that can work where humans work, instead of being confined to a fixed pedestal or tracks.

Modular end-effectors: At the end of each arm, DIGIT can equip different grippers or tools depending on the job. Agility provides standard suction grippers (useful for picking up plastic totes or boxes by their flat surfaces) and other swappable end-effectors agilityrobotics.com. The latest version of DIGIT unveiled in 2025 includes more robust, flexible limbs and wrist joints, expanding the range of motion and angles from which it can grasp items agilityrobotics.com. For example, it can now grab tote bins from higher or lower positions than before, and from different angles, making it easier to integrate into existing workflows. This modular design means companies can customize DIGIT’s “hands” to suit specific tasks – whether it’s lifting cartons, operating simple machinery, or even using warehouse tools.

Under the Hood: Software & Cloud Integration

Agility Robotics has developed a full software ecosystem around DIGIT to make it as autonomous and easy to deploy as possible. Each robot runs on-board software for perception, mapping, and movement. Notably, Agility uses reinforcement learning in simulation (via NVIDIA’s Isaac platform) to train some of DIGIT’s behaviors agilityrobotics.com – essentially letting a virtual DIGIT practice tasks and refine its neural network controls before those skills are uploaded to the real robot. This helps with complex maneuvers like balancing while picking from a shelf, or coordinating leg and arm motions smoothly.

In everyday operation, a DIGIT robot connects to Agility’s cloud-based platform called “Arc.” Arc serves as a fleet management and integration hub for the robots agilityrobotics.com agilityrobotics.com. Through Arc, human supervisors can coordinate multiple DIGITs, set up missions (like “go to aisle 3 and retrieve tote #12”), and monitor robot status. Arc also links DIGIT with other warehouse systems – for example, it can dispatch an Autonomous Mobile Robot (AMR) like a floor drone to ferry items that DIGIT has picked agilityrobotics.com. Agility emphasizes that humanoids and AMRs work best together, each playing to their strengths: DIGIT handles complex manipulation in human-centric spaces, while wheeled robots handle long-distance hauling agilityrobotics.com technologymagazine.com. In fact, at ProMat 2025 Agility demoed Digit working seamlessly with MiR and Zebra AMRs, using Arc to call an AMR when it needed to send off a picked tote agilityrobotics.com. This kind of systems integration is crucial in modern “smart” warehouses – DIGIT isn’t a standalone gadget, but part of a broader automated workflow.

Arc also provides over-the-air updates and a growing “Skill Library” for DIGIT agilityrobotics.com. Much like an app store, clients can download new robot skills as Agility develops them – for instance, a new palletizing skill to let DIGIT stack boxes on a pallet, or a parcel induction skill to unload packages from a conveyor. Agility’s vision is to build an “app store for labor,” where once you have the robot, you can continually expand what it can do via software improvements agilityrobotics.com. Some of the demonstrated skills so far include tote loading/unloading, order picking, box palletizing, and even tasks like recycling operations agilityrobotics.com. This software-centric approach means DIGIT can improve over time and adapt to different jobs without hardware changes.

User interface: For human operators, Agility provides tools to interact with and supervise the robots. There’s a wireless tablet pendant that an operator can use to manually guide DIGIT or move it between workstations safely agilityrobotics.com agilityrobotics.com. The newest DIGIT model has front-and-back status displays – small screens on its chest and back that show indicators like Wi-Fi connectivity, battery level, or error alerts agilityrobotics.com. This at-a-glance info helps workers nearby see what the robot is doing (or if it needs assistance). There are also multiple emergency stop buttons on the robot (often on the back panel) to immediately halt it if something goes wrong agilityrobotics.com agilityrobotics.com. Agility’s engineers have implemented a host of safety features (from Category 1 safe-stop functions that smoothly stop motion agilityrobotics.com, to industrial safety PLCs and FailSafe-over-EtherCAT communications agilityrobotics.com) to ensure DIGIT meets emerging robotics safety standards. While safety regulations for free-roaming bipedal robots are still evolving, Agility says it’s proactively equipping DIGIT to be “cooperatively safe” to eventually work alongside people on the floor agilityrobotics.com agilityrobotics.com.

Commercial Availability and Pricing

Unlike many humanoid robots that remain lab prototypes, DIGIT is commercially available today – but you won’t find it on Amazon for personal purchase. Agility’s clients are mostly large firms in logistics, e-commerce, and manufacturing who are looking to automate repetitive, physically demanding tasks. To make adoption easier, Agility Robotics offers two models for deploying DIGIT: a subscription-like Robots-as-a-Service (RaaS) model, or a direct purchase of a robot fleet agilityrobotics.com agilityrobotics.com.

Under the RaaS model, a company essentially hires DIGIT like an employee – paying a monthly or hourly fee that includes the robot itself, all necessary software, maintenance, and support. Peggy Johnson (Agility’s CEO) has publicly indicated the cost is about $30 per hour per robot for the full package reemanrobot.com. At first glance that might sound high, but it covers a machine that (in theory) can work day and night with minimal breaks. Johnson noted that at $30/hr, a customer’s fully-burdened cost breaks even in under two years compared to a human worker doing similar work reemanrobot.com. In other words, the ROI (return on investment) < 2 years, assuming the robot is utilized effectively. This calculus is based on a “fully loaded” human labor cost of around $30/hr (including benefits) in warehouse roles therobotreport.com. If DIGIT can take over tasks and work more hours, it can save money long-term – without replacing humans entirely, it’s pitched as handling the dull, injury-prone tasks and augmenting the workforce.

For companies that prefer capital expenditure, Agility also allows outright purchase of DIGIT units (exact pricing is not public, likely on the order of low-to-mid six figures per robot). Purchased robots still require a software subscription for the Arc platform and ongoing support, similar to how enterprise software is sold agilityrobotics.com. Agility advertises that with either model, clients can scale the number of robots up or down and always have access to the latest upgrades and skills agilityrobotics.com agilityrobotics.com. This flexibility is important since humanoid robots are a new type of automation – customers often start with a small pilot fleet, evaluate performance, then expand deployment if all goes well.

What does $30/hour cover? Essentially, everything needed to operate DIGIT. The robot itself runs on a rechargeable battery, and Agility provides autonomous charging docks so the robot can park itself and recharge when needed agilityrobotics.com. (The current DIGIT can operate up to 4 hours continuously per charge under typical workloads agilityrobotics.com, and Agility ensures you have enough robots or battery swaps to cover 24/7 operation if required.) The RaaS fee also covers cloud services (Arc platform use), software updates, any replacement parts or repairs, and support services such as remote monitoring of the robot’s status agilityrobotics.com agilityrobotics.com. In essence, Agility’s pitch is that you avoid the large upfront cost and complexity of maintaining a cutting-edge humanoid robot; instead you pay as you go, and Agility takes care of keeping DIGIT running and improving. This model also aligns incentives – Agility needs the robot to perform well for the customer to renew the contract, so they have a stake in ongoing success.

It’s worth noting that Agility is not alone in using RaaS for humanoids. Many robotics firms expect that initially, humanoid robots will be too costly or too rapidly evolving for straight sales, so subscriptions make sense. Figure AI, for example, also plans to offer its humanoid on a service model initially, and even Tesla’s Elon Musk has floated the idea of “Tesla Bots” working for an hourly wage equivalent. As the technology matures and manufacturing scales up, we could see sticker prices for humanoids fall dramatically (Agility’s own long-term goal is to make units affordable enough to buy en masse). In fact, analysts at Goldman Sachs noted a sharp 40% drop in component costs for humanoid robots in the past year, bringing typical unit manufacturing cost into the $30,000–$150,000 range depending on complexity goldmansachs.com. In the coming years, if a humanoid robot’s cost truly approaches that of a car, we may see more direct sales. For now, though, hiring a DIGIT as a service is the common path.

Current Use Cases and Industry Adoption

DIGIT’s design is general-purpose, but Agility Robotics has zeroed in on warehouse and logistics applications as the beachhead market. These environments have high demand for automation (due to labor shortages and e-commerce growth) and are structured enough for robots, yet still built for humans – perfect for a humanoid. Here are some of the real-world deployments and pilots demonstrating what DIGIT can do:

Tote Handling at GXO/Spanx: In 2024, Agility announced a milestone deal with GXO Logistics, a major third-party logistics provider, following a successful pilot techcrunch.com. At a GXO-run Spanx distribution center in Georgia, a fleet of DIGIT robots is now moving plastic tote bins of clothing. The workflow involves six autonomous mobile carts (from a company called 6 River Systems) that ferry totes around; when they arrive at a conveyor station, DIGIT picks up the totes from the AMR and places them onto the conveyor for the next stage of processing reemanrobot.com. The robots can handle both empty and loaded totes, reaching both low and high positions on the carts reemanrobot.com. This deployment is modest in scale (“a small fleet” was mentioned reemanrobot.com), but it’s significant as one of the first real revenue-generating humanoid robot operations. GXO’s confidence is noteworthy, as they also trialed another humanoid (Apptronik’s Apollo) around the same time techcrunch.com, indicating these logistics players are actively evaluating robot labor. The Spanx site deployment proved that DIGIT can reliably perform repetitive pick-and-place tasks for hours, working in tandem with other robots and existing conveyor systems.
E-Commerce Warehouse Trials: Agility has kept specific customer names under wraps, but it has hinted at deployments in retail and e-commerce warehouses beyond the GXO case agilityrobotics.com. These likely involve similar tasks – for example, unloading items from conveyor to carts, sorting packages, or fetching products in a fulfillment center. One high-profile partner is Amazon, which not only invested in Agility but also quietly piloted DIGIT in at least one Amazon fulfillment center in 2023 techcrunch.com. GeekWire journalists saw Digit prototypes testing “tote consolidation” at an Amazon facility near Seattle – meaning the robots organized storage bins after inventory was removed geekwire.com. Amazon has not publicly announced any large-scale deployment as of 2025 (their own in-house robots like Sparrow handle some tasks), but the testing signals strong interest. If Amazon eventually rolls out humanoids broadly, it would be a massive validation of the technology.
Manufacturing Support: Humanoids can also assist in factory settings. Agility markets DIGIT for light assembly, material handling in manufacturing lines, and internal logistics (e.g. moving parts between workcells) agilityrobotics.com. One known customer is Ford Motor Company, which partnered with Agility a few years back to experiment with last-meter package delivery using an earlier DIGIT model (famously, a DIGIT walked out of a driverless Ford van to deliver a box in a 2020 demo). While that specific use case (outdoor delivery) isn’t a current focus, it demonstrated DIGIT’s potential in handling objects in semi-structured environments. In 2025, Agility is more focused on factory logistics – for instance, car part handling or working alongside assembly line robots. Rival startup Figure AI announced a deal with BMW to use humanoids in auto manufacturing iotworldtoday.com, and Agility likely has similar undisclosed projects underway. The manufacturing domain may see humanoids loading machine tools, stocking parts bins, or doing quality inspections – tasks that require mobility and a degree of dexterity but not necessarily high payload.
Research and Innovation Sites: A handful of DIGIT units have gone to research labs and innovation centers in previous years (Agility sold early units to academic partners). However, as of late 2025, Agility’s emphasis is on enterprise deployments rather than one-off sales to researchers. They want DIGITs proving themselves in real jobs, not just in lab experiments. Still, some universities and corporate R&D centers likely use DIGIT as a platform to explore humanoid applications, contributing to its improvement.

Importantly, Agility claims that DIGIT is working in OSHA-regulated industrial environments, but currently with separation from human workers for safety reemanrobot.com. At the Spanx warehouse, for example, the DIGITs operate in a sectioned-off zone transferring totes between robots and conveyors, rather than walking freely among people. This controlled deployment is by design – until safety standards and trust in humanoids mature, most companies will start by using them in fenced or clearly defined areas. Over the next 1–2 years, Agility expects to clear the path for “functional safety” certification so that DIGITs can be truly collaborative robots (cobots) working in close proximity to humans reemanrobot.com reemanrobot.com. Achieving that would unlock far more use cases, from assisting human pickers on warehouse floors to stocking shelves in retail stores during normal operations.

The current adopters underscore a broader trend: warehousing and logistics are leading the early humanoid adoption curve reemanrobot.com. In 2023–25, several companies in this sector (logistics 3PLs, carmakers, electronics manufacturers) have unveiled pilot programs with humanoid robots. It’s a period of trial and error, where everyone is assessing what these robots do well and where they need improvement. As one robotics commentator put it, “the industry has been all promises and pilots – but now actual robots are starting to do meaningful work” techcrunch.com. Agility’s DIGIT deployments are often highlighted as evidence that the long-awaited age of humanoid helpers is finally beginning.

Latest News & Developments (as of 2025)

The past year (2024–2025) has been eventful for Agility Robotics and DIGIT, marked by rapid progress on technology and several “firsts” in the industry:

Upgrades Unveiled at ProMat 2025: In March 2025, Agility took the stage at ProMat (a major supply chain tech show) to introduce new features for DIGIT agilityrobotics.com. Key improvements focused on endurance, safety, and integration. The battery runtime was extended to ~4 hours of continuous operation, plus the robot can now autonomously dock at a charging station when it’s low agilityrobotics.com. This is crucial for scaling fleets – robots can recharge themselves in between tasks without manual battery swaps. Agility also revealed more robust manufacturing processes for DIGIT (hinting that they’ve streamlined production to crank out units faster) agilityrobotics.com. On the safety side, the newest DIGIT has multiple new layers of protection: a Category 1 stop capability for quick halting, an integrated Safety PLC computer monitoring its motions, wireless emergency stop on the teach pendant, and fail-safe communication links (FSoE) that meet high industrial safety standards agilityrobotics.com agilityrobotics.com. These are the kind of features required for certifying robots to work near humans. Melonee Wise, Agility’s Chief Product Officer, noted that these upgrades are paving the way for humans and Digit “to one day work side by side.” technologymagazine.com The theme was “cooperative safety”, reflecting the company’s strategy to go above and beyond minimal requirements so that customers and regulators trust humanoids on the floor agilityrobotics.com.
Integration with Other Robots: Another announcement at ProMat was Agility’s Arc platform now integrating with Autonomous Mobile Robots (AMRs) in live deployments agilityrobotics.com agilityrobotics.com. Agility demonstrated Digit working with MiR and Zebra Technologies’ mobile robots – effectively showing that a humanoid can be part of a broader automated workflow. For instance, Digit can call an AMR over, load it with items, and send it on its way, all orchestrated through cloud software agilityrobotics.com. This development addresses the “islands of automation” issue warehouses face agilityrobotics.com: various robots or automated systems that don’t talk to each other. By linking humanoids and wheeled robots, Agility wants to maximize overall efficiency – humans, humanoids, and AMRs each doing what they’re best at. This also showcases Digit’s higher-level autonomy: it’s not only performing physical tasks, but also making decisions like a logistics coordinator (e.g., dispatch robot “A” to take item “B” over to station “C”) agilityrobotics.com. Successfully pulling this off in a customer setting (they cited the GXO Atlanta site as an example agilityrobotics.com) is a notable innovation in 2025.
New Use Cases Demoed: On the ProMat show floor, Agility demonstrated that the upgraded Digit can handle a wider variety of warehouse jobs. Besides the tote picking we’ve seen, they showed Digit doing palletizing and depalletizing (stacking boxes onto pallets and vice versa), loading and unloading items from AMRs, manipulating items on flow racks and carts, and working an “automated putwall” (sorting station) agilityrobotics.com. These demos are important because they hint at the “Skill Library” that Agility is building. Each use case – say depalletizing – represents a set of software behaviors and possibly hardware tweaks. By proving them out in demos (and likely in trials with partners), Agility is expanding what tasks Digit can perform in a real warehouse. One by one, Digit is learning the fundamental chores of distribution centers, which could eventually allow a single robot type to do many roles that currently require separate fixed machinery or a lot of human labor.
Leadership and Funding News: Agility Robotics itself has grown and changed. In March 2024, Peggy Johnson (former Microsoft exec and ex-CEO of Magic Leap) took the helm as CEO techcrunch.com geekwire.com, bringing big-tech management experience to help scale production. Under her leadership, Agility has leaned even more into the ROI-driven narrative – focusing on Digit as a solution for today’s labor gaps, not just a cool research project techcrunch.com. The company reportedly began raising a large Series C funding round (~$400 million) in early 2025, which would value Agility at over $1.7 billion geekwire.com. Investors like SoftBank and WP Global were rumored to join, according to media reports geekwire.com. This influx of capital is earmarked to mass-produce Digit and develop next-gen tech. It follows a $150M Series B in 2022 (which notably included the Amazon Industrial Innovation Fund) agilityrobotics.com agilityrobotics.com. With the new funding, Agility is doubling staff and expanding its facilities – including the new Salem, OR “RoboFab” factory which was announced in late 2023 and began coming online in 2024 geekwire.com. RoboFab is touted as the world’s first humanoid robot factory, projected to eventually build over 10,000 Digits per year geekwire.com. By late 2025, Agility says it’s on track with that factory and expects to hit a 10k annual unit run rate within a couple of years reemanrobot.com. For comparison, only a few dozen Digit units existed in 2022; this shows how aggressively they are scaling up.
Next-Gen Digit on the Horizon: Agility has also teased the next iteration of Digit hardware. Peggy Johnson revealed that a new version of Digit was slated for release in fall 2024 with significant improvements reemanrobot.com. These include the aforementioned higher payload (50 lbs), longer battery life (potentially approaching 8+ hours with better batteries) reemanrobot.com, and general performance upgrades. By late 2025, it’s likely this next-gen model is what Agility is delivering to customers. Each version brings Digit closer to being a fully “enterprise-ready” humanoid – meaning more durable, easier to maintain, and with any early kinks worked out. The company is also continuously updating Digit’s software to add capabilities. For instance, Johnson mentioned working toward a “4-to-1 or even 10-to-1” utilization ratio in the future reemanrobot.com – implying that with better batteries and charging, one robot could do the work of 4 (or 10) over a given period by minimizing downtime. That kind of improvement would dramatically improve the economics for customers.
Public Perception and Policy: 2025 has also seen the conversation around humanoid robots shift from pure fascination to practical considerations. Agility Robotics often stresses that Digit “takes tasks, not jobs,” positioning it as a tool to handle menial labor amid workforce shortages agilityrobotics.com. This messaging is likely in response to concerns about robots displacing human workers. Notably, the Forbes article “The Robot That Takes Tasks, Not Jobs” (Aug 2025) highlighted Agility’s view that humanoids can fill essential gaps without mass layoffs, by doing work people don’t want or where there aren’t enough hands. Agility’s CEO has talked about focusing on solving real problems (like high turnover roles in warehouses) rather than chasing sci-fi visions. Meanwhile, organizations like OSHA and industry bodies are in the early stages of developing safety standards specific to bipedal robots agilityrobotics.com. Agility is directly involved in these conversations and often demonstrates an almost hyper-compliance approach (e.g. adding more safety features than currently required) to make Digit welcome in regulated workplaces. By late 2025, we’re seeing more public and regulatory acceptance of humanoids as potentially useful industrial machines, not just viral video fodder.

All told, Agility Robotics in late 2025 is a company transitioning from startup mode to scaled operations. Digit has evolved from a prototype to a product, and each news update shows it inching further into the mainstream of automation. The next big milestones to watch for will be: larger fleet deployments (do any warehouses order, say, 50 or 100 Digits?), entry into new domains (retail store trials? delivery pilots revived?), and how competitors respond – which brings us to the competitive landscape.

How Does DIGIT Stack Up Against Other Humanoid Robots?

The race to build useful humanoid robots has heated up fast, with a mix of startups and tech giants each pursuing their own designs. Here’s a look at DIGIT vs. some notable humanoid peers and how each is positioned:

Tesla Optimus – Ambitious Vision, Prototypes in Progress: Tesla shook the tech world by announcing its “Optimus” humanoid robot project in 2021, aiming to leverage Tesla’s AI and manufacturing might to create a general-purpose bipedal robot. As of 2025, Optimus exists as a prototype that Tesla has shown performing basic tasks (picking up objects, walking slowly, identifying items). Elon Musk has laid out an extremely bold roadmap: internally building all the tech (including high-dexterity hands capable of fine manipulation), ramping to thousands of units in 2025, and eventually millions annually with costs under $20k per robot at scale humanoidroboticstechnology.com humanoidroboticstechnology.com. In a late 2024 briefing, Musk even claimed Optimus could be “the majority of Tesla’s long-term value,” projecting staggering revenue potential humanoidroboticstechnology.com humanoidroboticstechnology.com. However, these projections remain aspirational – Optimus in 2025 is not deployed at any customer (initial units are being tested in Tesla’s own car factories doing simple tasks like carrying parts) humanoidroboticstechnology.com. Where Digit has a head start is real-world hardening: Optimus is still in a guarded development phase, working mostly in controlled environments, whereas Digit has been exposed to messy warehouse conditions and iterated upon. Tesla’s advantage could come from scale and AI: they are applying their self-driving AI expertise (which requires massive neural network training) to Optimus, and claim that the AI training compute for Optimus exceeds that for Tesla vehicles by 10x humanoidroboticstechnology.com. If Tesla hits its milestones, Optimus v2 in 2025 and v3 by 2026 might quickly narrow the gap. Tesla’s vision is Optimus eventually being a general-purpose home and industry robot, but near-term they’re targeting their own factories as the first use-case (much like Digit’s focus on warehouses). For now, Agility’s Digit enjoys first-mover advantage in commercialization, but Tesla’s project bears close watching given the company’s resources. It’s a classic agility vs. scale scenario: Agility Robotics is smaller and nimble, Tesla is slower today but aiming very big.
Figure AI’s Figure 01/02 – Stealthy Contender with Big Backers: Figure AI is a Silicon Valley startup (founded 2022) that in many ways is pursuing the same goal as Agility: a versatile humanoid for labor. Their humanoid, initially called Figure 01, was revealed in 2023 as a 5+ foot tall biped somewhat similar in form to Digit (though with a head and two eyes more reminiscent of a human face). Figure has attracted heavyweight funding – investors include OpenAI, Microsoft, and Nvidia – and in 2025 they announced plans to ship 100,000 humanoid robots over the next 4 years iotworldtoday.com iotworldtoday.com. How? CEO Brett Adcock says Figure is focusing on a couple of deep-pocketed pilot customers and will scale within those accounts (one known partner is BMW, which signed on to test humanoids in car manufacturing) iotworldtoday.com. Figure is taking a very AI-centric approach: they’re training end-to-end neural network “brains” for their robot to learn tasks autonomously iotworldtoday.com. Adcock even noted that writing manual heuristics for complex tasks is infeasible – implying they rely heavily on machine learning to teach the robot what to do iotworldtoday.com. In August 2024, Figure unveiled a refined prototype Figure 02 with improved hardware and AI, and talked about starting “alpha” tests of humanoids in real home settings by late 2025 techcrunch.com. Compared to Digit, Figure’s robots are at a slightly earlier stage (no known live deployment yet outside of test labs), but they are ramping up quickly. Both Agility and Figure share some common elements – for instance, both integrate Nvidia tech (Figure uses Nvidia for AI computing too) and both emphasize general-purpose utility. Figure’s differentiation might be in long-term vision: they explicitly mention eventual home use, whereas Agility has stayed laser-focused on industry. It’s conceivable that within a couple of years, Figure’s humanoid will be vying for the same clients as Digit. In the meantime, Agility can tout real performance data from Digit, while Figure works to prove its concepts. The competitive landscape could see Agility vs. Figure as a Coke vs. Pepsi of humanoid startups, each racing to secure marquee customers. Notably, both are collaborating with auto manufacturers (Agility with Ford in R&D, Figure with BMW) and large tech (Agility with Amazon, Figure with Microsoft/OpenAI support).
Boston Dynamics Atlas – The Athletic Superstar (Not for Sale): No discussion of humanoids is complete without Boston Dynamics’ Atlas, the robot that wowed the world with backflips and parkour. Atlas is a bipedal humanoid developed purely as an R&D platform to push the boundaries of agility and control. It’s extremely advanced in dynamic movement – able to jump between platforms, balance on a narrow beam, and even perform synchronized routines – feats far beyond anything Digit is today capable of. However, Atlas is not a commercial product; it’s essentially a testbed to develop technology that might trickle down into future robots. Boston Dynamics has famously avoided commercializing Atlas due to cost and practicality – Atlas units reportedly cost hundreds of thousands if not millions each, and require expert teams to operate. Instead, Boston Dynamics’ commercial strategy shifted to other robots (like Spot and Stretch, see below). So, while Atlas and Digit are both humanoid in form, they’re not exactly head-to-head competitors in the market. Atlas demonstrates the pinnacle of balance and locomotion – for example, it inspired the world by showing a humanoid could run and jump with grace. Agility Robotics has more modestly demonstrated Digit carefully walking and carrying things, which is arguably more immediately useful even if less acrobatic. There’s a bit of an arms race element here: Atlas shows what’s technically possible (e.g., leaping onto a platform to reach a high shelf), and companies like Agility no doubt watch those videos closely. In 2023, Boston Dynamics even showed Atlas doing some lab prototype “work” tasks like picking up a tool bag and tossing it to a worker – hinting they are researching gripper hands and manipulation humanoidroboticstechnology.com. But until Boston Dynamics decides to turn Atlas (or a derivative) into a product, Digit has no competition from Atlas in the field. Atlas’s influence is more indirect, raising public expectations and providing a pool of innovation that others can learn from.
Boston Dynamics Stretch – Specialized Warehouse Workhorse: Unlike Atlas, Stretch is a product and one that operates in the same domain as Digit – warehouses. However, Stretch looks nothing like a humanoid: it’s essentially a large wheeled base with an extendable arm and smart gripper. It was designed to do one thing really well – unloading trucks and stacking boxes onto pallets or conveyors. Stretch’s arm has a suction gripper that can grab boxes up to ~50 lbs, and its mobile base allows it to move up and down a trailer, reaching high shelves with ease. Boston Dynamics began shipping Stretch to warehouses in 2022–2023 for automating trailer unload, which is a brutal manual job in logistics. In comparing Stretch and Digit, we see two philosophies: purpose-built vs. general-purpose. Stretch is extremely efficient for the specific task of moving boxed inventory; it can likely unload a truck faster and for longer than a humanoid could. But it can’t do much else – it’s not going to walk to another part of the warehouse or pick a tote from a shelf, for example. Digit, on the other hand, can perform a wider variety of tasks but might do each individual task slower than a specialized machine. For a warehouse operator, the question is whether to use a fleet of specialized robots each for different jobs (one type for truck unload, another for order picking, etc.) or to use a flexible fleet of humanoids that can be reassigned to different tasks as needed. Agility Robotics argues the latter brings more flexibility and easier integration into human-centric workflows agilityrobotics.com agilityrobotics.com. In practice, we may see Digit and Stretch not directly competing but rather coexisting: one could imagine a warehouse where Stretch unloads pallets at the dock, then a Digit takes individual items from those pallets to sort onto shelves. Still, they will inevitably be compared on ROI. Stretch has the credibility of Boston Dynamics’ engineering and likely has a hefty price tag plus maintenance needs, whereas Digit offers a human-like form that might work with existing infrastructure (e.g., using a normal elevator to go between floors, something Stretch can’t do). It’s early to say which approach wins out; some industry voices think the future will involve a mix of form factors. Agility will be keen to prove that a humanoid can match the throughput of more specialized bots by virtue of improving rapidly in capability.
Apptronik Apollo – Newcomer with Modular Design: Apptronik is another startup to watch. Spun out of NASA research, they revealed Apollo, a human-sized humanoid, in 2023. Apollo stands about 5’8” and 160 lbs, very close to Digit’s size, and can lift up to 55 lbs – actually exceeding Digit’s current payload apptronik.com. Apollo is billed as a “general purpose” humanoid for mass manufacturing, with emphasis on ease of production and safety apptronik.com. Uniquely, Apptronik partnered with Jabil (a large contract manufacturer) to scale production of Apollo – a savvy move to leverage existing assembly lines humanoidroboticstechnology.com. They also entered a trial with GXO Logistics (yes, the same GXO that uses Digit) to test Apollo in warehouse tasks youtube.com. Apollo hasn’t been deployed commercially yet, but it’s on a fast track: at CES 2025 it was demonstrated performing basic box-handling tasks, and the company has raised significant funding (reportedly a $70M Series B in 2023, with a valuation around $350M) medium.com. Apollo’s design is described as “friendly” (it has a screen with animated eyes) and safety-focused, much like Digit’s cooperative approach apptronik.com. The competitive angle here is that multiple startups (Agility, Figure, Apptronik, 1X, etc.) are converging on similar product specs – human-sized, ~50 lb payload, battery-powered bipeds for industrial work. Each will differentiate on execution and partnerships. Agility has the lead in deployment and a mature software stack; Figure has the glitziest investors and AI-first strategy; Apptronik is leveraging manufacturing know-how. It will be a vigorous competition, but also these players share a common goal of legitimizing humanoid robots in the market. Notably, GXO and other early adopters are hedging bets by trying out multiple robots in parallel techcrunch.com. Whichever performs best will get scaled. Agility hopes that Digit’s track record – being “first with robots actually doing real work” as Peggy Johnson touts techcrunch.com – gives it the edge.
Others and Future Entries: The list doesn’t stop there. 1X (formerly Halodi) in Norway is working on a humanoid (called NEO) with backing from OpenAI as well. Sanctuary AI in Canada built a humanoid model (Phoenix) aimed more at dexterous hand tasks with teleoperation. Xiaomi’s CyberOne, Ubtech’s Walker, Toyota’s T-HR3, and others show that companies around the globe are experimenting with humanoid form factors for various niches (from service robots to elderly care). Even Honda and Hyundai (which owns Boston Dynamics) have hinted at next-gen humanoid projects, building on their legacy with Asimo and such. The competitive landscape in 2025 is thus a mix of startup agility and big-tech muscle. No single humanoid has “won” yet – in fact, we haven’t seen large-scale deployments of any make or model so far. But Agility’s Digit is often cited as a frontrunner because it has real customer usage and a clear short-term application focus technologymagazine.com. Tesla’s Optimus garners enormous attention due to Tesla’s influence (and the sheer scale of Musk’s plans), but it remains to be seen if they can execute as fast as promised.

One thing all players agree on: the potential market is enormous if they can crack the technology. There’s a sense of a coming paradigm shift – a “Cambrian explosion” of humanoid robots – possibly within this decade. The notion is that once one or two use cases are proven profitable, adoption could accelerate rapidly, and economies of scale will kick in to make these robots cheaper and more ubiquitous.

The Competitive Landscape and Market Outlook

Humanoid robots are no longer sci-fi fantasy; they’re emerging as the next big bet in tech. By late 2025, what was once a field of a few research labs has turned into a crowded industry race, and Agility Robotics with DIGIT has staked a claim as an early leader. But how large is this opportunity, really, and what trajectory do experts foresee for the market?

Analysts and major investors are increasingly bullish on humanoid robots. Goldman Sachs Research estimated in 2023 that the total addressable market for humanoid robots could reach $6 billion by 2035 – but just a year later, in 2024, they revised that dramatically upward to $38 billion by 2035 goldmansachs.com. In their report, Goldman cited faster-than-expected progress in AI (especially in generalizable, self-learning models) and rapid cost declines in robot components as reasons for the jump goldmansachs.com goldmansachs.com. They pointed out that many critical parts (sensors, actuators, batteries) are now being produced cheaper and better thanks to investment and supply chain expansion for electric vehicles and automation broadly goldmansachs.com. This cost reduction – roughly 40% drop in unit cost in one recent year goldmansachs.com – could bring humanoids to market sooner and make them economically viable for more applications. Goldman’s base case forecast is around 250,000 humanoid units shipped per year by 2030 (nearly all industrial at first), and then consumer humanoids picking up in the following decade goldmansachs.com.

Looking further ahead, Morgan Stanley released a report in 2025 dubbing humanoids a potential “$5 trillion market by 2050” morganstanley.com. That figure sounds astronomical, but it includes not just robot sales but an entire ecosystem – maintenance, software, services – akin to how the automobile market’s value includes fueling, parts, and so on. Morgan Stanley analysts predicted that by 2050 there could be 1 billion humanoid robots working in the world, with the vast majority in commercial roles morganstanley.com. They expect adoption to really accelerate in the 2030s once the tech is more mature and societies have adjusted regulations and acceptance of robot coworkers morganstanley.com. Interestingly, they note that currently China is investing heavily and could lead in humanoid development thanks to strong government support and manufacturing prowess morganstanley.com. This could foreshadow competition not just among companies but among nations in the robotics space.

From an industry standpoint, 2025 appears to be Year One of the humanoid commercial era. Agility’s Digit achieving paid deployments is a watershed that others are touting as well – for instance, Figure AI’s founder proudly announcing their first commercial deals, or 1X delivering units to customers. It’s reminiscent of the early days of personal computers or smartphones: different prototypes emerging, lots of hype, and uncertainty about which use case will break through first. There is also an element of hype vs. reality to manage. Some skeptics point out that today’s humanoids are still relatively slow, can only handle controlled tasks, and require significant human oversight or troubleshooting. They warn that over-promising could lead to a backlash if robots don’t live up to lofty expectations (some remember the “AI winters” of the past). However, the consensus in the tech and manufacturing community is that labor demographics and economics are driving a real need for these solutions. Populations are aging in many countries and companies struggle to fill manual jobs – robots like Digit are seen as timely innovations to take up the slack, augmenting human workers rather than outright replacing them agilityrobotics.com.

Agility Robotics finds itself in a favorable position amid this landscape. By focusing on a clear market (warehouse work) and steadily improving Digit’s capabilities in response to real customer feedback, they have built a credible narrative that “humanoids are here and working now.” Their challenge will be maintaining that lead as deep-pocketed competitors ramp up. It’s telling that Agility is rapidly raising capital and scaling manufacturing – they likely anticipate a wave of demand and also the need to stay ahead on tech. The company’s partnership strategy (e.g. working with logistics giants, integrating with AMRs, aligning with standards efforts) indicates they want to make Digit as easy to adopt as possible, effectively setting a benchmark for humanoid utility in industry.

For the broader market, one can foresee a scenario where by the late 2020s, humanoid robots become a fairly common sight in warehouses, factories, and maybe retail stores. They might still be slower or less “smart” than human workers, but if they can reliably perform simple tasks 24/7, the economics will drive adoption. Each successful deployment will further validate the concept, attracting more investment and accelerating development – a virtuous cycle. On the flip side, any major failures (say a high-profile crash or safety incident, or a big project quietly being shelved) could pump the brakes temporarily.

In summary, DIGIT stands at the forefront of a nascent market that is poised for exponential growth. Its early deployments have raised the bar for what a humanoid robot can do outside the lab techcrunch.com. But this is just the starting line of a long race. Agility Robotics faces competition from the likes of Tesla, Figure, Boston Dynamics, and others – each bringing unique strengths. The next few years will reveal whether DIGIT can maintain its lead as more humanoids step into the ring. If the optimistic projections hold true, the humanoid robotics sector will transform from a curiosity into a multi-billion dollar industry solving real-world problems at scale. And Agility’s DIGIT will have played a historic role as one of the pioneers that proved the concept – showing that a walking, human-shaped robot can take on tedious tasks, work safely and reliably, and help write a new chapter in automation history.

Sources: Agility Robotics press releases and tech briefs agilityrobotics.com technologymagazine.com; TechCrunch and Forbes interviews with Agility’s leadership techcrunch.com agilityrobotics.com; GeekWire and IoT World Today reports on industry developments geekwire.com iotworldtoday.com; Goldman Sachs and Morgan Stanley market research goldmansachs.com morganstanley.com; and others as cited throughout.