iFlytek's Liu Qingfeng: Domestic computing power still about two generations behind the global top, but the gap is narrowing rapidly

iFlytek's Liu Qingfeng: Domestic computing power still has about a two-generation gap with the global top, but it is narrowing rapidly

Quick Tech Download the App | Exclusive First Look October 22, 2026, 17:06:33 From Henan Updated October 22, 2026, 17:14:41

Quick Tech News, October 22: Today, iFlytek Chairman Liu Qingfeng stated at a public event that artificial intelligence is a great technological transformation of extraordinary weight in the course of human history, and that it is also a long-distance race that requires the entire industry to unite and advance steadily; it is impossible to secure final industrial victory through short-term, speculative, explosive investment.

Liu Qingfeng mentioned that, benchmarked against the world's top level as a whole, domestic computing power still has a technological gap of about two generations.

But this gap is narrowing rapidly at a speed far exceeding outside expectations; it will soon be shortened to one generation, and before long it will be narrowed to only half a generation. In the end, we will certainly catch up with the world's top technology and achieve the goal of running neck and neck and even leading in some fields.

He also candidly acknowledged that, in the entire domestic technology ecosystem, there are not many enterprises that are truly settling down and investing core resources in deep cultivation at the underlying training layer of large models, but he believes that in the future there will certainly be more and more industry participants proactively plunging into R&D on underlying large-model training technology and into related work such as building a richer full-chain application ecosystem.

Liu Qingfeng further emphasized that China simply cannot afford to lose on this core track of artificial intelligence. Relying on the country's huge talent pool, rich application scenarios, and complete industrial supporting systems, we will certainly be able to secure a unique core competitiveness in the global AI industry and hold our own unique voice.

The core prerequisite for achieving this goal is for the entire industry to unite on a fully independent and controllable technology platform and jointly tackle the bottleneck points in underlying technologies.

Earlier, Liu Qingfeng had publicly offered his own judgment that Huawei, in the dimension of intelligent inference computing power, has provided the global market with a second fully viable technological choice besides Nvidia. The actual performance of its related self-developed computing power chips is already at the same level as mainstream international products, and it has demonstrated quite outstanding adaptation advantages in multiple dimensions, including adaptation to the domestic computing power system, large-model training optimization, and industry application in deployment scenarios.

Many AI industry practitioners also say that the domestic large-model industry has now passed the lively early stage of blossoming everywhere, and that what all vendors will compete on next is the long-term accumulation of underlying computing power and independent ecosystems. The entire industry pulling together to tackle core technologies is indeed the only viable path for us to gain a voice in global AI competition.

"Special statement: The above content (including videos, pictures, and audio if any) is uploaded and published by users of 'Dafeng Hao,' a self-media platform under Phoenix Net. This platform only provides information storage space services.

Notice: The content above (including the videos, pictures and audios if any) is uploaded and posted by the user of Dafeng Hao, which is a social media platform and merely provides information storage space services."