On August 19, during the World Robot Conference (WRC), Huixi Intelligent released its Huixi Embodied product series. The launch focused on chips, modules, software development tools, and engineering deployment — the robots' movements themselves were only one part of the story.
From Chip to Workstation
The R1 PRO is a fused big-brain/small-brain SoC for embodied scenarios: it puts CPU, NPU, GPU, and real-time interfaces on a single chip, aiming to close the loop of perception, reasoning, decision-making, and motion control on-die. Official specifications include:
- 375 TOPS INT8 AI compute
- 16-core Cortex-A78AE CPU
- 1 kHz real-time control
- 204.8 GB/s LPDDR5 bandwidth
- Up to 22 camera inputs
- Ethernet, CAN FD, PCIe, USB, and other interfaces common in robotics
Above the chips sit the RCM 400 and RCM 500 compute modules. RCM 400 pairs with R1 PRO, offering 16GB/32GB/64GB LPDDR5; RCM 500 pairs with R1 MAX, offering 32GB/64GB/128GB. Both use a 699-pin interface and the RISE / Robotics SDK. The modules absorb repeated design work — memory, UFS storage, power management, clocking, and high-speed signals — so robotics companies don't have to start from high-layer PCB design every time.
What WRC Says About the Industry
First-day reporting at WRC noted that more than 200 embodied-intelligence exhibitors attended this year, with 300+ physical robots running on site; exhibits shifted from backflips toward guided tours, logistics handling, production-line assembly, and inspection. The conference also held a "Procurement Day" for the first time. These are media observations, not orders for any specific company, but they explain why Huixi chose this moment to launch a full-stack platform: customers have started asking about delivery timelines, interface stability, and continuous operating costs. The questions have changed.
Huixi also released an AI-augmented development kit, attempting to let agents participate in model deployment, performance profiling, operator development, and tuning. The company's stated goal: compress deployment cycles for the same algorithm from 2 weeks to 2 days. This claim deserves a careful read. The 2-weeks-to-2-days figure is a launch-event target, not independently measured third-party data. What actually needs verification is under what models, sensors, and power constraints the reduction holds, whether operators must be rewritten, and whether on-site failure rates actually drop.
One SoC Isn't Enough
TOPS is a peak metric and doesn't directly translate into a robot's success rate at grasping, sorting, or assembly. The hard constraints of embodied systems include sensor time synchronization, control-loop jitter, memory bandwidth, thermals, interface latency, and software migration costs. By putting these functions on one chip, R1 PRO's value lies in reducing multi-chip communication and board-level integration complexity; whether it enables skill reuse across different robot bodies still depends on whether the SDK, model formats, and training data are truly shared.
This is also where the news differs from the earlier narrative of "one brain across bipedal, wheeled, and heavy-load bodies." Galbot (银河通用) emphasizes generalization at the model and body level; Huixi is platformizing the compute foundation and engineering interfaces. The two are adjacent on the industry chain, not substitutes.
For embodied intelligence to cross the "demo works" threshold, the next gate comes down to plain questions: how much code changes when one algorithm moves to another robot, how many manual restarts a week of continuous running requires, and who maintains low-level drivers during production-line upkeep. Chip launches tend to hide these answers in a spec sheet — WRC's Procurement Day pushes them back into the open.
Sources and verifiable links
1. 36Kr: Huixi Intelligent releases the Huixi Embodied series (2026-08-19) https://36kr.com/p/3946324792556672 2. 36Kr: WRC 2026 day-one observations (2026-08-20) https://36kr.com/p/3946697386808967