Collaborative Robotics (company)
Last reviewed
Sources
18 citations
Review status
Source-backed
Revision
v1 · 1,895 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Sources
18 citations
Review status
Source-backed
Revision
v1 · 1,895 words
Add missing citations, update stale details, or suggest a clearer explanation.
Collaborative Robotics, which operates under the brand Cobot, is an American robotics company founded in 2022 and headquartered in Santa Clara, California, with a second office in Seattle, Washington. It builds a practical wheeled autonomous mobile manipulator called Proxie that is designed to move carts, totes, and boxes alongside people in warehouses, hospitals, manufacturing plants, and other workplaces. The company was started by Brad Porter, a former vice president and distinguished engineer at Amazon Robotics who later served as chief technology officer of Scale AI. Collaborative Robotics is notable in part for the form factor it deliberately avoids: rather than chasing a humanoid robot, Porter has publicly argued that a human-shaped machine is the wrong tool for most material-handling work, and the company has positioned Proxie as a more pragmatic alternative. By April 2024 it had raised more than 140 million dollars across seed, Series A, and Series B rounds, the last of which valued the company at over 500 million dollars.
This article is about the company Collaborative Robotics. For the general category of human-safe industrial robots, see collaborative robot.
Collaborative Robotics was founded in 2022 by Brad Porter, who serves as chief executive officer. Porter spent roughly 13 years at Amazon, where he rose to vice president and distinguished engineer of Amazon Robotics and oversaw the deployment of hundreds of thousands of robots across the company's fulfillment network, including sorters, palletizers, drones, sidewalk robots, and automated storage-and-retrieval systems. He left Amazon in August 2020 to become the first chief technology officer of Scale AI, the data-labeling and AI-infrastructure company, a role he held for just under two years before starting Collaborative Robotics. Porter holds bachelor's and master's degrees in computer science from MIT, and his earlier career included stints at Netscape and the voice-technology company Tellme Networks.
The company emerged from stealth in June 2022 with a five-person team that Porter planned to grow to 20 to 25 staff, predominantly engineers. Over time the team came to include robotics and AI specialists drawn from Amazon, Apple, Meta, Google, Microsoft, NASA, and Waymo. From the outset the actual design of the robot was kept under wraps; in its early funding announcements the company described only a "novel cobot" without showing the hardware, and trade press referred to it as a mysterious mobile manipulator until the product was revealed in late 2024.
A defining feature of Collaborative Robotics is its rejection of the humanoid robot form factor that many of its venture-backed peers have pursued. Porter has framed the company's target as "something much more pragmatic in the middle" between a general-purpose two-legged android and a low-profile autonomous mobile robot that can only shuttle goods along the floor. He has argued that humanoids tend to be expensive, lack the strength to handle heavy loads, suffer from limited battery life, and require especially complex AI to balance and walk, which makes them poorly suited to the bulk of real warehouse and logistics tasks. The company instead bet on a wheeled platform that could carry and move heavy carts immediately and reliably while still working safely around humans.
Collaborative Robotics raised capital quickly across three rounds in its first two years. The figures below are drawn from the company's own announcements and from venture databases and trade press.
| Round | Date | Amount | Lead investor | Notes |
|---|---|---|---|---|
| Seed | June 2022 | 10 million USD | Neo | With Khosla Ventures, Calibrate Ventures, and 1984 Ventures |
| Series A | July 2023 | 30 million USD | Sequoia Capital | With Khosla Ventures and Mayo Clinic; launched the Cobot Flywheel program |
| Series B | April 2024 | 100 million USD | General Catalyst | Valued the company at over 500 million USD; total raised exceeded 140 million USD |
The 30 million dollar Series A, announced on July 26, 2023 and led by Sequoia Capital, brought total funding past 40 million dollars and was earmarked for scaling early field deployments and manufacturing. Alongside the round, the company introduced its Cobot Flywheel program, a structured pilot effort aimed at companies in biotechnology, healthcare, and logistics; the Mayo Clinic adopted the program as part of its collaboration with the startup.
The 100 million dollar Series B, announced on April 10, 2024, was led by General Catalyst and joined by Bison Ventures, Industry Ventures, and Lux Capital, with existing backers Sequoia Capital, Khosla Ventures, Mayo Clinic, Neo, 1984 Ventures, MVP Ventures, and Calibrate Ventures also participating. Crunchbase News reported that the round valued the company at more than 500 million dollars and ranked among the largest robotics financings of 2024, a year in which investment in robotics startups surged. As part of the Series B, General Catalyst managing director Paul Kwan and Sequoia partner Alfred Lin joined the board, and Teresa Carlson, a former head of Amazon Web Services' worldwide public sector business, joined as an advisor. The company said it would use the capital to expand its team and advance commercial deployments.
The company publicly unveiled its robot, named Proxie, on November 20, 2024. Proxie is a mobile manipulator roughly the height of a person, built to take on repetitive material-handling work such as moving carts and transporting totes and boxes so that human workers can focus on higher-value tasks. The company markets it as a practical collaborative robot, or cobot, rather than a general-purpose android, and emphasizes that it is designed to be trusted and safe to work near. At launch the company said Proxie was already operating in production settings across logistics, healthcare, and manufacturing.
Proxie is neither a humanoid robot nor simply a robot arm bolted onto a standard mobile base. Its mobility platform, which the company calls Glide 360, uses a four-wheel swerve-drive architecture that gives it pseudo-holonomic motion: it can move in any direction and slide sideways, for example to tuck into a row of shelving, which the company says provides smoother and more flexible movement than differential-drive or mecanum-wheel designs. Instead of a multi-fingered hand, Proxie's standard configuration uses a vertical lifting column with a cart-interfacing gripper, branded Flex Grasp, that lets it engage, lift, and move existing carts and containers. The robot can move carts carrying payloads of up to about 1,500 pounds (roughly 680 kilograms). A six-degree-of-freedom collaborative robot arm has been described as an optional future configuration rather than a standard feature.
For perception, Proxie uses a system the company calls Scout Sense, which captures the surrounding environment from roughly human eye level so the robot perceives doorways, carts, and people from a vantage point similar to a worker's. The company also highlights an "empathetic" industrial design, with visual cues and stability-focused hardware intended to make the robot's intentions legible to nearby people and to build trust in human-robot collaboration.
Proxie's autonomy stack is built on NVIDIA hardware and software. According to NVIDIA and the company, each robot runs on two NVIDIA Jetson AGX Orin modules plus a Jetson Nano: one Orin handles simultaneous localization and mapping (SLAM) and camera-based scene perception, the second Orin manages human-machine interaction including the use of large language models to interpret voice commands, and the Nano handles lower-level motor control and power management. Scout Sense perception runs on NVIDIA Isaac accelerated libraries and AI models on the Orin modules. The company has described a proprietary planning layer it calls the Auditable Collaboration and Planning Framework (ACoP), which converts human speech and text instructions into executable robot actions in a way meant to be inspectable.
The team relies heavily on simulation to deploy quickly. Using NVIDIA Isaac Sim, Collaborative Robotics builds digital twins of customer facilities, populated with pre-labeled "SimReady" assets such as pallets and dock doors, and validates a deployment in simulation before sending robots to a site. The company has said this simulation-first approach let it stand up its first Proxie units at customer locations in under six weeks from the completion of assembly. Like a number of other contemporary robots, Proxie draws on generative AI and vision-language models so that it can interpret its surroundings and natural-language instructions rather than relying solely on hard-coded routines, which places it in the broader category of AI-driven robotics systems that combine perception, language, and ai agents-style task planning.
Collaborative Robotics has run pilots and early deployments with several large enterprises in logistics and healthcare. The company has named Maersk, the Mayo Clinic, Moderna, Owens & Minor, and Tampa General Hospital among its first customers. The Mayo Clinic is both an investor and a deployment partner, having adopted the Cobot Flywheel program. Tampa General Hospital has worked with the company to move carts and reduce elevator congestion in a clinical setting. The shipping and logistics company Maersk has tested the robots in warehouse and material-handling operations.
By December 2024, around the time of Proxie's public launch, the company reported that roughly 30 robots were operating across customer sites, that they had logged more than 5,000 hours of operation, moved about 16,000 carts, and traveled more than 621 miles in total. Proxie was recognized in The Robot Report's RBR50 awards for 2025. As of early 2026, third-party databases listed the company's headcount in the range of about 150 to 160 employees. The most recent disclosed financing as of mid-2026 remained the April 2024 Series B; no later round had been publicly announced.