CtrlK
| RoboForce TITAN | |
|---|---|
| General Information | |
| Manufacturer | RoboForce |
| Type | Industrial mobile manipulator |
| Year unveiled | 2025 |
| Status | Pilot deployments / pre-commercial |
| Founders | Leo Ma (CEO), Calvin Zhou (VP Engineering) |
| Headquarters | Milpitas, California, United States |
| Website | roboforce.ai |
TITAN is an AI-powered industrial mobile manipulator developed by RoboForce, an American robotics startup headquartered in Milpitas, California. Designed for demanding outdoor and heavy-duty applications, TITAN combines dual-arm manipulation with autonomous navigation capabilities to perform physically intensive tasks in industries such as utility-scale solar installation, mining, manufacturing, logistics, and data center construction.[1][2] The robot is powered by RoboForce's proprietary RF-Net 3D foundation model, which provides advanced spatial understanding and enables operation in complex, unstructured environments with millimeter-level precision.[3]
RoboForce positions TITAN as the centerpiece of its "Robo-Labor" strategy, which aims to deploy autonomous robotic workers for jobs classified as dull, dirty, and dangerous. The company has attracted significant attention from the robotics industry and was highlighted by NVIDIA CEO Jensen Huang during the GTC 2025 keynote for its contributions to Physical AI.[4] As of March 2026, RoboForce has raised a total of $67 million in venture funding and has secured more than 11,000 robot orders through letters of intent across six industries globally.[5]
RoboForce was founded in June 2023 by Leo Ma and Calvin Zhou. The company is headquartered in Milpitas, California, in the heart of Silicon Valley.[6]
Leo Ma, who serves as Founder and CEO, holds a Master's degree in Computer Science from Carnegie Mellon University and completed executive education at the Wharton School of the University of Pennsylvania. Before founding RoboForce, Ma co-founded Cyngn (NASDAQ: CYN), an autonomous industrial vehicle company, where he served as VP of Engineering for six years and helped bring the company public in October 2021. Earlier in his career, he worked as a Software Architect for Autonomous Driving at Baidu USA and served as a Guest Lecturer and Software Engineer at Carnegie Mellon University.[7][8] Ma has cited a lifelong interest in production automation, inspired by his family's manufacturing background, and has analyzed automation practices firsthand in over 200 factories throughout his career.[8]
Calvin Zhou, Co-Founder and VP of Engineering, also holds a Master's degree in Computer Science from Carnegie Mellon University and a Bachelor's degree in Industrial Engineering with a minor in Computer Science from Tsinghua University. Zhou previously worked at Cruise (the General Motors-backed autonomous vehicle company) and at Cyngn. He also co-founded nLink, which built the world's first construction drilling robot, a project later acquired by Hilti, where Zhou played a key role in developing the Jaibot drilling robot.[9]
RoboForce's engineering team draws from leading institutions and technology companies. Team members include PhDs and experienced professionals from Carnegie Mellon University Robotics, the University of Michigan Robotics program, Amazon Robotics, Google, Waymo, Cruise, Tesla Robotics, ABB, and Apple.[5] The team's combined expertise spans control systems, robotics, perception, mechanical engineering, and autonomous systems.
RoboForce has raised $67 million in total venture funding across multiple rounds.
| Date | Round | Amount | Key investors | Total raised |
|---|---|---|---|---|
| January 2025 | Early-stage | $10 million | Myron Scholes (Nobel Laureate), Gary Rieschel (Qiming Ventures / SBVC), Carnegie Mellon University | $10 million |
| May 2025 | Follow-on | $5 million | New and existing investors | $15 million |
| March 2026 | Growth round (oversubscribed) | $52 million | YZi Labs (lead), Jerry Yang (Yahoo! co-founder), plus existing investors | $67 million |
The $10 million early-stage round was announced at CES 2025 in January and attracted high-profile individual backers including Myron Scholes, a Nobel Laureate in Economics, and Gary Rieschel, founding partner of Qiming Ventures and co-founder of SoftBank Venture Capital.[6][10] Carnegie Mellon University, Leo Ma's alma mater, also participated as an institutional investor.[6]
In May 2025, coinciding with the public launch of TITAN, RoboForce announced an additional $5 million from both new and existing investors, bringing total capital to $15 million.[1]
The most significant funding milestone came in March 2026, when RoboForce secured $52 million in an oversubscribed round led by YZi Labs, a $10 billion investment fund. New investor Jerry Yang, co-founder and former CEO of Yahoo, also participated. Ella Zhang, Managing Partner and Head of YZi Labs, joined RoboForce's board of directors as part of the investment.[5][11] The company stated the capital would be used to advance its next-generation robotics foundation model, scale manufacturing and supply chain operations, and drive commercial deployment and revenue growth.[5]
On June 24, 2025, coinciding with the company's second anniversary, the World Economic Forum (WEF) designated RoboForce as a 2025 Technology Pioneer. The company was recognized as the only U.S. AI robotics company in that year's cohort of 100 Technology Pioneers. The WEF Technology Pioneer program honors startups whose innovations transform advanced manufacturing, supply chains, and the Fourth Industrial Revolution. Past alumni of the program include Google, Twitter, and Airbnb.[12]
TITAN is classified as a mobile manipulator rather than a traditional bipedal humanoid robot. The robot stands approximately 2.1 meters (6 feet 11 inches) tall, making it one of the larger platforms in the industrial robotics space.[2][3] Rather than walking on two legs, TITAN uses a mobile base that is available in interchangeable wheeled and tracked configurations, with additional configurations planned for the future.[1] This design prioritizes ruggedness and terrain adaptability for industrial worksites over human-like locomotion.
RoboForce designs its hardware in-house because off-the-shelf components, including sensors, could not meet the reliability requirements for the harsh conditions in which TITAN is intended to operate.[2] The robot features a highly modular hardware architecture with interchangeable end-effectors, allowing flexible deployment across various environments and task types.[1][13]
TITAN is equipped with dual arms that provide a 40 kg (88 lb) payload capacity and 1,100 mm (43.3 in) of arm reach across the full workspace.[1][2] The manipulation system achieves approximately 1 mm (0.04 in) positioning precision, enabling fine-grained manipulation tasks required in industrial settings.[1]
The robot delivers five fundamental manipulation primitives that RoboForce considers essential for most industrial tasks:[1][2]
| Primitive | Description |
|---|---|
| Pick | Grasping and lifting objects from surfaces or storage |
| Place | Positioning and releasing objects at target locations |
| Press | Applying controlled force for assembly or fastening operations |
| Twist | Rotational manipulation for screwing, turning, or alignment tasks |
| Connect | Joining components such as cables, connectors, or structural elements |
RoboForce asserts that with these five base capabilities, TITAN can handle the vast majority of industrial manipulation tasks encountered in the field.[2]
TITAN operates on battery power and delivers an 8-hour continuous production runtime, matching a standard industrial work shift.[1][2] This endurance figure is a key selling point for the platform, as it allows the robot to work a full shift without requiring a battery swap or extended charging break.
The robot supports WiFi and 5G connectivity for remote monitoring, fleet management, and data transmission back to RoboForce's cloud infrastructure for model training and improvement.[14]
| Category | Specification | Value |
|---|---|---|
| Physical | Height | 2,100 mm (6 ft 11 in) |
| Physical | Form factor | Mobile manipulator (dual-arm) |
| Manipulation | Payload capacity (dual-arm) | 40 kg (88 lb) |
| Manipulation | Arm reach | 1,100 mm (43.3 in) |
| Manipulation | Positioning precision | ~1 mm (0.04 in) |
| Manipulation | Primitives | Pick, Place, Press, Twist, Connect |
| Mobility | Base options | Wheeled, tracked (interchangeable) |
| Power | Battery life (continuous) | 8 hours (480 minutes) |
| Computing | Onboard processor | NVIDIA Jetson Thor |
| AI | Foundation model | RF-Net (3D foundation model) |
| Connectivity | Wireless | WiFi, 5G |
| Software | License type | Closed source |
| Autonomy | Level | Level 3 (conditional autonomy) |
| Actuators | Type | Electric |
| End effectors | Design | Interchangeable / modular |
TITAN is powered by RF-Net, RoboForce's proprietary 3D foundation model for robotic spatial understanding. RF-Net provides the robot with advanced 3D spatial comprehension, allowing it to perceive, interpret, and interact with complex outdoor scenes and adapt to unstructured layouts in real time.[3][4] The model is central to what RoboForce calls its "Domain Intelligence" approach: a tight co-design of AI software and robotic hardware optimized specifically for industrial environments.[1][2]
Unlike robots that rely primarily on 2D camera images or pre-mapped environments, RF-Net builds rich three-dimensional representations of the workspace, enabling TITAN to handle the variability and unpredictability of outdoor industrial sites. This capability allows the robot to generalize learned skills across various tasks and environments without extensive reprogramming.[3]
A core element of RoboForce's technical strategy is its Physical AI data flywheel. Each deployed TITAN robot generates operational field data during its work shifts. This real-world data is transmitted back to the cloud, where it feeds into the training pipeline for RF-Net and other models. The improved models are then deployed back to the robot fleet, creating a continuous cycle of improvement.[5][11]
RoboForce describes this as a compounding system: "The more you use it, the smarter it gets." As the fleet grows and accumulates more operational hours across diverse environments, the foundation model becomes progressively more capable and reliable.[11] The data flywheel combines real-world fleet data with high-fidelity simulation to create a closed-loop reinforcement learning environment.
RoboForce has built its AI and computing stack in deep collaboration with NVIDIA, leveraging multiple components of NVIDIA's robotics platform:[5]
| NVIDIA Technology | Role in TITAN |
|---|---|
| Jetson Thor | Edge computing for onboard inference and real-time decision-making |
| Isaac Sim | High-fidelity robotics simulation for testing and validation |
| Isaac Lab | Open framework for robot learning and policy training |
| Cosmos | Synthetic data generation to augment real-world training data |
| OSMO | Cloud-to-edge orchestration for managing compute across fleet |
This technology partnership was publicly acknowledged when NVIDIA CEO Jensen Huang highlighted RoboForce during his GTC 2025 keynote in Washington, D.C. on October 28, 2025. Huang featured TITAN among a select group of companies demonstrating breakthroughs in Physical AI, endorsing the broader thesis that "AI is transforming factories into intelligent thinking machines."[4]
TITAN operates at Level 3 conditional autonomy, meaning the robot can handle routine operations and some unexpected situations autonomously but may require human supervision in certain edge cases.[14] This places TITAN in a practical middle ground: capable enough to work independently through full shifts, but with human oversight available when the robot encounters novel scenarios outside its training distribution.
RoboForce targets industries characterized by physically demanding, hazardous, or repetitive labor where worker shortages and high attrition rates constrain growth. The company has identified six primary target sectors for TITAN deployment.
Solar energy installation is TITAN's lead application and the first sector where pilot deployments have been conducted. RoboForce debuted its Robo-Labor solution publicly at Intersolar 2025 in San Diego, California, in February 2025, where the company demonstrated an all-terrain robot performing solar panel installation tasks from ground to grid.[15]
According to the Solar Energy Industries Association, worker shortages and rising labor costs are constraining growth in the solar sector, with over 4,400 GW of solar capacity projected globally by 2030.[11] In 2024, an estimated 53 gigawatts of U.S. solar projects were delayed due to labor scarcity alone.[11] RoboForce claims its robots can perform solar installation labor at three times the efficiency of human workers and at one-third of the average U.S. labor cost for equivalent work.[15]
Mining operations involve exposure to extreme temperatures, dust, hazardous materials, and physically demanding tasks in remote locations. TITAN's rugged design, all-terrain mobility (particularly its tracked base configuration), and 8-hour operational endurance make it suited for mining applications where human labor is both scarce and subject to significant safety risks.[1][2]
TITAN's five manipulation primitives (pick, place, press, twist, connect) align closely with common manufacturing operations. The robot's 40 kg payload capacity allows it to handle heavy components, while its 1 mm precision supports tasks requiring accurate positioning.[1]
With the rapid expansion of data center infrastructure driven by artificial intelligence workloads, RoboForce sees growing demand for robotic labor in data center construction and maintenance tasks.[5]
Logistics hubs and shipping facilities involve repetitive material handling, loading, and unloading tasks. TITAN's dual-arm configuration and payload capacity position it for these applications.[5]
RoboForce has identified the space industry as a target market, though specific deployment plans for this sector have not been publicly detailed. The robot's ability to operate autonomously in unstructured and extreme environments aligns conceptually with space-related applications.[6]
RoboForce uses the term "Robo-Labor" to describe its approach to deploying autonomous robots as a workforce solution for industrial tasks. The concept centers on the idea that certain categories of work, those that are dull, dirty, and dangerous, should no longer require human workers, particularly given growing global labor shortages in these sectors.[5][6]
The company's value proposition emphasizes three key advantages:
| Factor | Claimed benefit |
|---|---|
| Cost | 66% cost savings: $300,000 in value delivered for $100,000 in cost |
| Speed of deployment | 4 weeks to deploy, compared to a typical 52-week hiring process (90% faster) |
| Reliability | 90% reduction in operational uncertainty by eliminating human labor management challenges |
CEO Leo Ma has stated: "Robo-Labor is essential for work that is dull, dirty, and dangerous. This problem centers on human workers' availability, cost, and safety."[5]
TITAN enters a rapidly growing market for industrial and humanoid robots. The broader competitive landscape includes a range of companies developing robots for industrial and commercial applications.
| Company | Robot | Approach |
|---|---|---|
| Tesla | Optimus | General-purpose humanoid; bipedal; factory and household focus |
| Figure | Figure 02 | General-purpose humanoid; BMW manufacturing partnership |
| Apptronik | Apollo | Humanoid for logistics and manufacturing |
| Agility Robotics | Digit | Warehouse-focused humanoid; Amazon pilot deployments |
| Sanctuary AI | Phoenix | General-purpose humanoid with carbon AI control system |
| Boston Dynamics | Atlas | Humanoid for industrial R&D and commercial pilots |
| RoboForce | TITAN | Industrial mobile manipulator; outdoor/harsh environment focus |
TITAN differentiates itself from many competitors by focusing specifically on rugged outdoor industrial environments rather than indoor factory floors or general-purpose household use. While companies like Tesla, Figure, and Agility Robotics have emphasized bipedal humanoid form factors, RoboForce has opted for a wheeled/tracked mobile base that prioritizes stability, payload capacity, and terrain adaptability over human-like locomotion.[2][3] This focused approach targets what the company sees as a more immediately addressable market: heavy industrial jobs that existing robots cannot perform and that human workers increasingly refuse to take.[13]
| Date | Milestone |
|---|---|
| June 2023 | RoboForce founded by Leo Ma and Calvin Zhou in California |
| January 2025 | $10 million early-stage funding announced at CES 2025 |
| February 2025 | Public debut of Robo-Labor at Intersolar 2025, San Diego |
| May 2025 | TITAN officially launched; $5 million additional funding (total: $15 million) |
| June 2025 | World Economic Forum designates RoboForce a 2025 Technology Pioneer |
| October 2025 | NVIDIA CEO Jensen Huang highlights TITAN at GTC 2025 keynote |
| March 2026 | $52 million funding round led by YZi Labs (total: $67 million) |
As of early 2026, RoboForce has reported securing more than 11,000 robot orders through letters of intent (LOIs) across multiple industries.[4][5] The company is actively collaborating with customers across six industries globally, with initial pilot deployments already active in the energy sector. Expansions to data centers, shipping, logistics, mining, and manufacturing are underway.[4]
The commercial launch of TITAN was originally planned for the end of 2025, following pilot programs with early customers that commenced earlier in the year.[15] With the $52 million funding round closed in March 2026, RoboForce has stated its focus is on converting pilot programs into full production deployments and establishing recurring revenue streams.[5]