China’s rising appetite for large models and AI-enabled services has pushed demand for compute into a new stratosphere, and one Shenzhen-based company is pitching a domestic solution. Pinggao Co. (品高股份), in close cooperation with Jiangyuan Technology, has unveiled a full-stack, “autonomous and controllable” AI system — branded the Pingyuan AI all-in-one — which the company says stitches together domestically produced chips, system-level hardware and software to reduce reliance on foreign suppliers.
The move is emblematic of a broader national priority. As AI use cases multiply across finance, manufacturing and public services, policymakers and industry executives are urging accelerated investment in indigenous compute chips, richer talent pipelines, and co‑ordinated industry-academia platforms. Pinggao’s management is calling for targeted measures during China’s upcoming 15th Five‑Year Plan (2026–2030): a dedicated innovation fund for AI compute chips, stronger R&D tax breaks, improved talent attraction and retention policies, and tighter intellectual property protections to foster supply‑chain collaboration.
Technically the company stresses hardware-software co‑design — from chip to rack-level compute orchestration — arguing that performance gains will come as much from integration and scheduling software as from the silicon itself. That argument responds to two market realities: first, the complexity and cost of developing cutting-edge GPUs and accelerators; second, the fact that system-level orchestration, datacenter integration and software toolchains often determine real-world AI performance more than raw chip spec sheets.
The pitch also has a geopolitical backdrop. Western export controls on advanced AI accelerators have elevated the stakes of domestic capabilities. For Chinese firms and policymakers, import substitution in critical compute hardware is as much about economic opportunity as it is about risk mitigation — safeguarding key AI applications from supply‑chain disruption and sanction risk.
Yet hurdles remain. High-performance AI silicon requires not only design expertise but advanced fabrication, packaging and a mature ecosystem of compilers, libraries and datacenter management tools. China has accelerated its investment in these areas, but gaps remain at the high end of process technology and in the commercial software stacks that make new chips broadly useful. Building a viable domestic alternative therefore demands sustained capital, talent and time, as well as pragmatic industry coordination.
For international markets, China’s push matters even if domestic challengers cannot immediately match the top-tier accelerators from the likes of NVIDIA. A credible domestic compute industry would reshape supply chains, alter pricing dynamics, and give Beijing greater leverage in setting the rules for AI deployment at scale. Whether Pinggao’s all-in-one system proves a technical or commercial breakthrough, its emergence is a clear signal that China intends to lay a homegrown foundation beneath its digital economy.