On August 20, at the 2026 World Robot Conference, just after the forum ended, GenstorAIge Vice President of Products – Li Zhen was surrounded by reporters from Beijing TV, Nanfang Daily, and Guangzhou Daily. He provided an in-depth interpretation on "how data flows through and how computing power is unleashed."

Photo: GenstorAIge Vice President of Products – Li Zhen interviewed by on-site media reporters

GenstorAIge's "Cloud-Edge-Device" Strategy Emerges

"For a single humanoid robot, the raw data stream from sensors like vision, depth, IMU, and joint encoders alone reaches GB/s levels. Data is generated, but computing power cannot keep up—this is the biggest disconnect in current edge-side AI." Li Zhen gave interviews to reporters from Beijing TV, Nanfang Daily, and Guangzhou Daily respectively.

Photo: GenstorAIge Vice President of Products – Li Zhen interviewed by on-site media reporters

He explained that GenstorAIge's breakthrough approach centers on opening up a complete "Cloud-Edge-Device" AI data link: the cloud handles large model training and massive data storage, the edge serves as a token factory and real-time inference hub, and the device side equips each unit with localized AI capabilities. This enables data to flow efficiently between cloud, edge, and device, rather than being bottlenecked at each layer.

"The core capability running through all three layers is compute-storage convergence." Li Zhen emphasized in the interview, "Others are making more expensive GPUs; GenstorAIge is making smarter storage. Solving the computing power bottleneck with storage—this is the key to the edge and device side."

Nanfang Daily & Guangzhou Daily Joint Interview on Embodied Intelligence – GenstorAIge's Second Battlefield at the Edge and Device

When reporters from Beijing TV and Guangzhou Daily pressed for details on embodied intelligence application scenarios, Li Zhen identified this as the most representative AI application direction for the edge and device side.

"The implementation of embodied intelligence cannot rely on single-point breakthroughs. Data generated at the device, processed at the edge, and trained in the cloud—this entire link must be opened up for robots to truly scale from 'piloting a few units' to 'deploying thousands.'"

He further explained GenstorAIge's core products for embodied intelligence device-side scenarios to the two reporters:

AI90 is GenstorAIge's device-side product for compute-storage convergence. Through intelligent P2P technology and intelligent KV Cache scheduling, it resolves GPU communication bottlenecks, making large model inference faster and more efficient on the device side.PT200Z, GenstorAIge's self-developed AI SSD, features TB-level massive capacity, high bandwidth, and low latency. Within the three-tier cloud-edge-device data architecture, it serves as high-performance storage support for cloud-based large model training and inference scenarios.

Photo: GenstorAIge Vice President of Products – Li Zhen interviewed by on-site media reporters

GenstorAIge's "Cloud-Edge-Device Strategy" to Make a Major Debut in September

Li Zhen's sharing sparked intense follow-up questions ranging from "how to break through the device-side memory wall" to "how to build a data closed loop for embodied intelligence." And this is just the prelude to a series of major moves by GenstorAIge—in September, GenstorAIge will officially release its complete cloud-edge-device collaborative solution.

"As data becomes the new oil of the robot era, GenstorAIge aims to enable the flow of data in the AI era—ensuring every bit of data leads to intelligence."

On September 17, GenstorAIge's AI Strategy Launch Event. Stay tuned.

｜About GenstorAIge:

GenstorAIge, with the mission of "Enabling Data to Lead to Intelligence," is a leading hard-tech company in China that uniquely integrates three layers: Storage, Connectivity, and Solutions. It offers a full-stack product portfolio including enterprise SSDs, PCIe Retimer/Redriver connectivity chips, and integrated AI training and inference solutions.