An in-depth look at 1X Technologies, the Norwegian-founded humanoid robotics company behind the NEO robot, tracing its origins as Halodi Robotics, its $240M in funding from NVIDIA and SoftBank, and its strategy of using home environments to train AGI-capable AI systems.
1X Technologies takes a practical approach to humanoid robotics, focusing on how robots can be useful in the home first, before anywhere else. Founded in 2014 as Halodi Robotics by Norwegian engineer Bernt Øivind Børnich, the company has grown from its Scandinavian roots into a global player. It now operates out of Palo Alto, with R&D hubs in Sunnyvale, California, and Moss, Norway.
The team’s core goal is to support an “abundant society” by developing safe, smart, and affordable humanoid robots. Its flagship model, NEO, is built for tasks like cleaning and caregiving - designed to help people with real needs at home.
1X Technologies is also using home environments to help train AI systems for broader capabilities. Homes are complex, unpredictable spaces, making them ideal training grounds for general-purpose intelligence. By placing robots in dynamic, human-centric spaces, 1X Technologies believes it can accelerate the development of AGI, positioning itself as a leader in both robotics and the broader quest for human-level intelligence.
1X Technologies began as Halodi Robotics, focusing on safe actuators and full-body control systems for industrial and healthcare robotics. In 2018, it introduced EVE, a humanoid robot on wheels designed for logistics, security, and medical environments.
By 2022, the company rebranded as 1X Technologies, pivoting toward legged humanoid robotics with a vision of deploying general-purpose humanoid robots in homes. This strategic shift was driven by Børnich’s belief that robots must live and learn alongside humans to achieve true intelligence, a philosophy that aligns with 1X Technologies’s home-first approach.
In 2025, 1X Technologies made headlines by acquiring Kind Humanoid, a Palo Alto-based startup founded by former Google robotics researcher Christoph Kohstall. This acquisition brought Kohstall’s expertise and the bipedal robot Mona into 1X Technologies’s fold, strengthening its Bay Area operations and accelerating its path to market. The move could be interpreted as a technology and acquihire. Bernt Bornich said that “Having Christoph join the 1X Technologies team here in the Bay will accelerate our path to a world full of humanoid robots.”
1X Technologies’s ambitious vision has attracted some of the biggest names in tech and venture capital.
In March 2023, 1X Technologies raised $23.5 million in a Series A2 round led by the OpenAI Startup Fund - its first investment in a humanoid robotics company. The move highlighted a shared focus on embodied AI and data-driven learning.
The round also saw participation from prominent investors such as Tiger Global, known for its bets on transformative tech; Sandwater, a Nordic VC firm with an eye on sustainable innovation; Alliance Ventures, the investment arm of Renault-Nissan-Mitsubishi; and Skagerak Capital, a longtime supporter of 1X Technologies with strong local ties in Norway.
The funding helped scale up production of 1X Technologies’s EVE robot and supported early development of its next-generation platform, NEO.
In 2024, 1X Technologies raised an impressive $200 million in a Series B round at an $850 million post-money valuation, led by EQT Ventures, a European firm renowned for backing scalable tech disruptors. The round attracted a who’s-who of tech giants, including NVIDIA, SoftBank Group, Samsung Electronics NEXT, Skagerak, Nistad Group, Jeff Bezos (through Bezos Expeditions), and Sandwater.
NVIDIA’s involvement is particularly noteworthy, as the chipmaker’s expertise in AI and computing complements 1X Technologies’s ambitions, with NEO frequently featured in demos alongside NVIDIA’s CEO, Jensen Huang, highlighting a growing partnership that leverages NVIDIA’s hardware to enhance NEO’s AI capabilities. This $240 million in total funding was all about refining its Redwood AI World Model, and solidifying its leadership in safe, affordable humanoid robotics for the homes.
NEO is 1X Technologies’s flagship product, it is a 66-pound, 5-foot-5-inch, tendon-driven humanoid purpose-built for intrinsic safety, making it ideally suited for integration into homes.
Launched as NEO Beta in August 2024 and succeeded by the advanced NEO Gamma in February 2025, NEO represents a monumental step toward practical, safe, and intelligent home robotics. NEO Gamma features a natural human gait, enhanced capabilities to squat, sit, and manipulate objects with precision, and is powered by 1X Technologies’s proprietary large language model (LLM) that translates natural conversations into actionable tasks.
Its hardware, significantly more reliable and 10 decibels quieter than its predecessor, incorporates new features like four microphones for superior audio capture and a three-speaker system for immersive voice interaction and sound effects.
These advancements, driven by 1X Technologies’s custom Revo1 servo motor, which delivers 5.5 times the torque-to-weight ratio of previous leading motors, ensure quiet, powerful, and efficient motion. The rapid hardware progress suggests that the upcoming NEO Delta could be poised for larger scale commercial deployments.
NEO’s initial focus is on household tasks such as caregiving, cleaning, and cooking, targeting the $12.5 trillion household labor market. Its bio-inspired, modular design enables 1X Technologies to adapt NEO for diverse verticals, including hospitality, retail, security, logistics, and eldercare, through software updates rather than expensive hardware redesigns.
Recent demonstrations highlight NEO’s progress in voice control, task chaining, and autonomous behavior, with 1X Technologies transitioning from human teleoperation for complex tasks to increasing AI-driven autonomy as its Redwood Vision-Language-Action (VLA) model matures. Posts on X reflect growing excitement, with Eric Jang, 1X Technologies’s Head of AI, revealing that NEO Gamma has been rigorously tested in employee homes for weeks, and select early adopters will soon receive units under nondisclosure agreements.
Eric Jang, who leads AI at 1X Technologies, believes that the company is on the path to becoming an AGI company, outside of robotics. The strategy revolves around collecting vast, diverse datasets from real-world tasks, unattainable in controlled settings like factories. Homes and public spaces, with their varied environments and challenges, are ideal for gathering the data essential to developing human-like intelligence. 1X Technologies seeks to amass a critical data trove to fuel this ambition.
Redwood, 1X Technologies’s Vision-Language-Action (VLA) model, drives NEO’s whole-body control, enabling precise task execution. Currently focused on immediate actions, the next iteration will tackle task planning. The model is quite compact which is an advantage for fast response time.
1X Technologies’s proprietary World Model (1XWM), an end-to-end neural network, is used to train Redwood by integrating vision, language, and motor control and is specifically designed to model the home environment. 1XWM goes beyond the NVIDIA’s Isaac GR00T N1, by simulating human behavior, and capturing critical real-life interactions NEO will encounter in homes. It also models household-specific elements like cloth and liquids, which are prevalent in domestic settings but less common in rigid environments like factories or warehouses.
Redwood’s strength lies in its ability to create a self-reinforcing AI flywheel. Each deployed NEO robot acts as a real-time data factory, continuously collecting manipulation data from real-world interactions, which 1X Technologies uses to refine and enhance Redwood’s capabilities. With over 10,000 hours of data already amassed, this feedback loop positions NEO as a versatile platform for tasks ranging from household chores like cleaning and caregiving to eventual expansion into hospitality, retail, logistics, and eldercare.
The model’s architecture supports rapid iteration, enabling 1X Technologies to optimize robot policies through efficient training-evaluation cycles, select the best training checkpoints, curate datasets for long-tail scenarios, and re-evaluate models to ensure robustness in production environments.
