Simplexity Robotics, founded in late July 2025, made its first public appearance and officially announced its financing.
Recently, Simplexity Robotics (至简动力) announced that it had completed five consecutive funding rounds within just six months, raising a total of RMB 2 billion and quickly becoming one of the youngest unicorns in the embodied intelligence sector.
The investor lineup is notably strong, bringing together several top-tier venture capital firms and strategic investors. Financial investors include Yuanjing Capital (元璟资本), BlueRun Ventures (蓝驰创投), HongShan (红杉中国), Legend Capital (君联资本), CAS Star (中科创星), and Gaorong Capital (高榕创投). Strategic investors include Tencent (腾讯) and Alibaba Group (阿里巴巴集团). The latest round was advised by Lighthouse Capital (光源资本) as the financial advisor.
It is rare in this sector for Tencent and Alibaba Group to participate in the same funding round. Public information also shows that most early investors—including Tencent, HongShan, BlueRun Ventures, and Legend Capital—have continued to increase their stakes across multiple rounds since the company’s first investment.
Five funding rounds in six months, RMB 2 billion raised, and repeated follow-on investments from existing shareholders have quickly drawn strong market attention. Some investors noted that Simplexity Robotics has demonstrated remarkable iteration speed over the past few months: almost every meeting with the team comes with new milestones achieved and new targets set.
In many ways, Simplexity Robotics is moving at “lightning speed,” demonstrating strong advantages in its technology roadmap, commercialization potential, and team execution capability while reinforcing investor confidence.
Simplexity Robotics
Simplexity Robotics (至简动力) is an innovative technology company focused on building high-value embodied intelligence products grounded in real-world scenarios. The company aims to empower future human development through a combination of high-ceiling unified models, efficient data closed loops, and highly reliable robotic hardware.
Its core philosophy, “Simple is Scalable,” runs through the company’s entire technology architecture and product design. Simplexity Robotics believes that scalable simplicity is the key to solving the inherent complexity of embodied intelligence.
Simplexity Robotics adopts a full-stack approach with in-house development across both software and hardware. Guided by the philosophy of “models defining the embodiment and software defining the hardware,” the company has built a streamlined technical framework centered on the “Four O’s”:
One Model
On Device
One Body
One Hour
This architecture enables a unified and efficient development paradigm for embodied intelligence systems.
At the model level, the company has developed an integrated architecture that combines a world model with Vision-Language-Action (VLA) capabilities. Built on a unified Transformer framework, the system jointly models and predicts language reasoning, visual semantics, 3D spatial structures, and robotic states. This design reduces manual engineering while significantly improving scalability.
Currently, Simplexity Robotics has introduced several core technologies:
LaST₀ Foundation Model
The LaST₀ foundation model integrates the world model’s physical-world understanding and prediction capabilities with the fast-and-slow reasoning paradigm of VLA systems. This architecture significantly improves robots’ ability to reason about dynamic physical environments, addressing the fundamental challenge of enabling robots to “think while acting.”
ManualVLA Long-Horizon Task Model
Built upon the LaST₀ foundation, ManualVLA focuses on enabling robots to understand and execute complex long-horizon tasks. Starting from a target state, the model can autonomously generate multimodal “operation manuals” similar to those used by humans, enabling robots to plan before acting.
The related research paper has been accepted by CVPR 2026.
TwinRL Real-World Reinforcement Learning Framework
Beyond generalization and reasoning capabilities provided by the base model, Simplexity Robotics also focuses on enabling robots to continuously learn and evolve in real-world environments.
The TwinRL framework expands the exploration space of real-world reinforcement learning through digital twins. In multiple tasks, robots can achieve 100% task success rates within less than 20 minutes of training in tabletop environments, addressing the challenge of enabling robots to continuously improve through real-world interaction.
Additional model research from Simplexity Robotics is expected to be released in the near future.
The company believes that solving the data closed-loop challenge in embodied intelligence requires several key principles: full-stack in-house development across both hardware and software, a model-defined embodiment, a unified general-purpose robot body, and strong confidence in the scaling effect of data.
Recently, Simplexity Robotics proposed a new robot learning paradigm summarized as “Human Data Is All You Need.” The approach has already been validated across various dexterous manipulation tasks, including both grippers and multi-finger dexterous hands.
During the pre-training stage, large-scale operational data is efficiently collected through human demonstrations, significantly improving the model’s generalization capability. In the downstream task stage, human demonstrations enable rapid collection of task-specific data, expanding the exploration space and improving execution accuracy. During post-training, humans can provide real-time guidance to participate in model refinement, allowing robots to achieve efficient online learning and continuous capability improvement.
At the same time, the company embeds additional computing power on-device to support edge-side deployment and training. Through a shadow mode, models can be tested and validated directly in real user scenarios while continuing to learn at the edge. This paradigm allows Simplexity Robotics to significantly improve data universality and reusability, forming a highly efficient closed-loop system covering data collection, training, testing, validation, and deployment.
When it comes to embodied intelligence products themselves, Simplexity Robotics maintains the same philosophy of radical simplicity—simple deployment, simple operation, and simple maintenance.
According to the company, such simplicity ultimately reflects a deeper principle: placing user value at the center of product design. Time is the most valuable cost, and users should not have to pay for unnecessary complexity. Only by creating value for users can a company create lasting value for itself.
At present, the company’s first-generation self-developed robotic embodiment has already entered small-batch production and begun proof-of-concept (PoC) validation.
System-Level Execution Capability
The pace of Simplexity Robotics’ development has moved in tandem with its rapid fundraising. From the arrival of its first employee to the launch of its first self-developed robotic prototype, the company took less than 45 days.
Today, the company has established a strategic presence across Beijing, Shanghai, and Suzhou, built joint laboratories with leading universities, and launched a global innovation center in Suzhou. At the same time, it continues to advance its core model research and translate those capabilities into product performance.
Simplexity Robotics has already completed the development of two generations of robotic embodiments designed for both enterprise and consumer markets, with its robotic systems entering small-batch production and full-scale PoC validation. In parallel, the company closed five funding rounds within just six months.
As CEO Jia Peng (贾鹏) puts it, “System capability is our differentiation.”
Mass production and large-scale deployment of embodied intelligence products require a system-level approach, integrating strategy, technology, branding, product development, organizational design, and commercialization.
The core team at Simplexity Robotics has previously operated in highly competitive markets characterized by rapid iteration and intense pressure. Having experienced the full journey from survival challenges to market breakthroughs, the team has validated its ability to scale—from 0 to 1 and from 1 to large-scale commercialization.
This experience has shaped a deeply embedded system-level operational capability. The team understands how to efficiently integrate resources, manage cost and execution pace, and build a lean, efficient organizational structurecapable of making and executing core decisions across the entire value chain.