DeepSeek CEO Reveals Huawei's AI Chips Are Only Two Years Behind Nvidia

The artificial intelligence landscape is witnessing a dramatic shift as major technology leaders evaluate hardware capabilities globally. Recently, the Chief Executive Officer of the prominent AI startup DeepSeek shared critical insights into the competitive standing of Chinese semiconductor manufacturing, noting that Huawei's advanced hardware is closing the gap with industry giants much faster than anticipated.

Bridging the Technological Gap in AI Hardware

During an investor-focused conference call, Liang Wenfeng discussed various market dynamics, including data annotation workflows, the advantages of advanced processing architectures, and China's rapid progress in domestic computing power. A major highlight of the discussion centered on how Huawei Ascend solutions are positioning themselves as viable alternatives to Nvidia hardware.

Liang emphasized that the technological delta is shrinking rapidly. While foreign hardware has traditionally dominated the high-performance computing sector, local innovations are scaling to meet heavy workloads. For more details on related developments, you can check out the DeepSeek Huawei integration report.

Resource Scaling and Economic Viability

Despite the narrow two-year performance gap, operational scaling requires careful resource management. According to DeepSeek leadership, achieving massive model training milestones requires deploying a larger quantity of local processors to match the output of single foreign units. Specifically, training massive parameters relies on scaling cluster sizes significantly.

Nevertheless, the exceptional cost-effectiveness of these domestic solutions ensures that local enterprises continue shifting their infrastructure investments toward homegrown silicon. This transition is actively reshaping supply chains and fostering a self-reliant technological ecosystem.

How far behind Nvidia are Huawei's AI chips according to DeepSeek?

According to DeepSeek CEO Liang Wenfeng, Huawei's Ascend artificial intelligence chips are estimated to be only about two years behind Nvidia's industry-leading technology.

Can Huawei's hardware replace Nvidia systems directly?

Yes, systems like the Huawei Atlas 950 SuperPoD can fully match the performance and pricing structure of advanced Nvidia configurations, though deploying them may require a higher ratio of processors per task.

What is the main challenge when using domestic Chinese AI processors?

The primary challenge lies in resource scaling, as large-scale model training requires a greater quantity of local graphics processing units to achieve parity with alternative architectures.

Why is demand for Ascend processors increasing?

Demand is surging primarily due to their exceptional cost-effectiveness and strong capabilities in handling intensive computational tasks efficiently within local markets.

🔎 As the artificial intelligence hardware ecosystem continues to evolve at a breakneck pace, the narrowing gap between global semiconductor leaders and domestic innovators highlights a profound transformation in technological self-reliance. With companies actively optimizing architectures and scaling local resources, the future of high-performance computing points toward a heavily diversified and competitive global market.