About QbitAI

QbitAI

QbitAI, published at qbitai.com and known in Chinese as Liangziwei, meaning Quantum Position, is a Beijing artificial intelligence news platform founded in 2017. It publishes fast AI industry news, large-model and robotics coverage, and produces widely cited Chinese AI talent and industry reports through its QbitAI think tank operation. Its bylines, presented in syndication as Quantum Position from the Aofei Temple, a playful reference to its Beijing newsroom, have become one of the most recognisable signatures in Chinese technology journalism, attached to stories that are routinely republished and quoted across the Chinese internet. Where its older peer Synced built its identity around bilingual research tracking, QbitAI built its identity around speed, proximity to China's AI companies and researchers, and an ability to turn a fast-moving technical field into vivid, shareable news without abandoning specialist knowledge. It arrived in 2017, as deep learning investment and talent competition were intensifying, and it was perfectly positioned for the large-model era that began several years later, when its daily coverage of model releases, startup financing, robotics and AI personalities became essential reading for China's technology sector.

Founded in Beijing in 2017

QbitAI was founded in 2017 in Beijing. By that year, the conditions for a dedicated AI news platform were fully formed. Chinese firms, finding AI personnel scarce at home, were already recruiting internationally, including in Silicon Valley, and by 2018 AI engineers in China could reportedly command higher salaries than their Silicon Valley counterparts. A May 2019 survey by China Youth Daily found that 78 percent of respondents were keen on working in the sector and almost as many were optimistic about its prospects, evidence of how completely AI had captured professional imagination in China. A new publication serving that world needed to do more than explain what a neural network was. Its readers were the researchers, engineers, founders and investors creating the talent shortage, and they needed a daily record of who had released what, who had raised money, who was hiring and whose claims were credible. QbitAI's name itself signals technical literacy: a qubit is the basic unit of quantum information, a reference its Chinese name renders as quantum position, and its branding assumes an audience comfortable with the vocabulary of advanced computing.

Fast news for the large-model era

Speed is QbitAI's defining editorial characteristic. Its model of coverage is visible in the stories bearing its byline across Chinese syndication: a 21-year-old founder's AI hardware product attracting more than 20 million views for a 99-dollar AI companion necklace; Mobvoi becoming the first AIGC stock on the Hong Kong exchange, closing its first day with a market value of 5.4 billion, its founder Li Zhifei's stake worth more than one billion; and Fei-Fei Li's World Labs exceeding a one billion dollar valuation in less than four months, having raised about 100 million dollars from investors including Andreessen Horowitz and Radical Ventures. Each story combines a hard news core, with names, figures and dates, with an explanation of why the development matters and a feel for the personalities involved. When the large-model wave transformed AI from a specialist sector into the central story of global technology, that style scaled naturally. Model releases from Chinese and foreign laboratories, price cuts by ByteDance, Tencent, Baidu and Alibaba following DeepSeek's low-cost breakthroughs, and the race to build AI agents and hardware all fell within the beat QbitAI had been covering since before generative AI had a mass audience.

Covering DeepSeek and China's model race

No episode better illustrates QbitAI's beat than the rise of DeepSeek. The Hangzhou company was founded in July 2023 by Liang Wenfeng, co-founder of the quantitative hedge fund High-Flyer, which remained its principal investor. It released its DeepSeek-V3 model and, in January 2025, an eponymous chatbot alongside the reasoning model DeepSeek-R1, released under the MIT License with performance comparable to OpenAI's GPT-4 and o1 class models. The company reported training V3 for about six million dollars, far less than the reported 100 million dollar cost of GPT-4 in 2023, using approximately one-tenth the computing power of Meta's comparable Llama 3.1 model, and it achieved those results while training under trade restrictions on AI chip exports to China. After R1's release, other Chinese technology giants, including ByteDance, Tencent, Baidu and Alibaba, cut the prices of their AI models, in what was described as a price war. For QbitAI's readers, such a story is simultaneously research news, business news and national industry news, and its coverage treats it as all three, following the technical methods, the corporate responses and the talent and investment consequences in continuing daily reporting rather than in a single announcement-day article.

The QbitAI think tank and its reports

QbitAI's most cited work beyond daily news comes from its research and think tank publishing. Its best-known early report was produced with International Data Corporation and released at the MEET 2020 Smart Future Conference in Beijing. Described as the first extensive study of the state of AI in China, the joint white paper found that although 73 percent of companies had implemented multiple AI applications, over 60 percent of respondents said implementation was difficult or extremely difficult, and 51.7 percent said the greatest challenge they faced was the lack of qualified personnel. The findings gave empirical shape to the talent war its news pages reported anecdotally, and they were reported internationally as evidence about the world's second-largest AI market. The think tank has continued in that vein during the generative AI period, including a China AIGC Advertising and Marketing Industry Panorama Report examining how generative AI and large models were changing the production factors, delivery effects and competitive landscape of advertising and marketing, analysing players from internet giants with their own models and dedicated research teams to marketing service providers building vertical models. Reports of that kind extend QbitAI from a newsroom into an industry research institution whose publications companies cite in their own planning.

AI talent as a continuing story

Talent is a thread running through QbitAI's identity, from the IDC white paper onward. The subject suits a Beijing AI publication because the city's universities, institutes and company laboratories form one of the world's densest concentrations of researchers, and because the movement of a prominent scientist can signal a company's strategy as clearly as a product launch. Recruitment data has repeatedly confirmed the story's importance: in spring 2019, a recruitment trend report found that six of the fifteen most in-demand skill sets in China were AI-related, with demand for internet-of-things engineers up 46 percent year on year and salaries in the field reaching nearly three times the national average. QbitAI's widely cited talent and industry reports, noted in its profile as a platform, draw on that environment, tracking where AI researchers work, how they move between academia and industry, and which fields attract them. For readers, the value is diagnostic. Publication counts and model benchmarks measure what AI systems can do; talent reporting measures whether the organisations behind them can keep doing it, and where the next generation of Chinese AI companies is likely to form.

Robotics and AI hardware

QbitAI's coverage extends from software models into robotics and AI hardware, the areas in which artificial intelligence becomes a physical product. Its reporting follows humanoid and embodied robots, AI devices and the startups building them, with the same combination of technical detail and founder narrative found in its model coverage. The AI necklace story is characteristic: a young founder who had built a widely visited COVID tracking website as a seventeen-year-old high school student raised 2.5 million dollars for a wearable AI companion company valued at 50 million dollars, and spent 1.8 million dollars acquiring its domain name. QbitAI's account carried those verifiable business facts while conveying why the product had become a social media phenomenon. Robotics and hardware matter increasingly to its beat because China's AI industry competes not only in cloud models but also in manufacturing, where its supply chains give embodied AI companies a distinctive advantage. A Beijing newsroom covering both the model laboratories and the hardware startups can therefore report the full stack of the industry, from a training method to the device in which a consumer finally meets it.

Voice, bylines and syndication

QbitAI's authorial voice is part of its brand. Its articles appear across Chinese platforms credited to Quantum Position and its WeChat public account, with individual writers signing from the Aofei Temple, a newsroom joke that regular readers recognise instantly. Individual bylined reporters recur across its coverage of major stories, giving the publication named, accountable authors rather than anonymous aggregation. Syndication through aggregators and partner sites carries those bylines far beyond qbitai.com, so that a QbitAI story about a Hong Kong listing, a Silicon Valley valuation or a Hangzhou model release may be encountered by readers who never visit the home site. That distribution pattern is typical of Chinese mobile-era media, in which a publication's influence travels through public accounts and republication as much as through its own domain. It also explains QbitAI's agenda-setting power. When its framing of a development, as the first AIGC stock, the Pinduoduo of AI or a valuation achieved in under four months, is repeated by other outlets, the publication has shaped the vocabulary in which the Chinese industry discusses itself.

Place among Chinese AI media

QbitAI operates in a field that includes Synced, founded in Beijing in 2014 with its Chinese service Jiqizhixin and English edition Synced Review, the AI channels of broader technology outlets such as Shenzhen's Leiphone, founded in 2011, and a large ecosystem of analyst firms, newsletters and company publications. Its position within that field rests on three assets: speed in daily industry news, a talent for narrative framing that makes technical stories travel, and a think tank whose reports are cited by companies and international media as data sources rather than commentary. The IDC white paper is the clearest example of the third asset, because its percentages on implementation difficulty and talent shortage entered the international record on Chinese AI. QbitAI is younger than Synced and narrower than the general technology press, but the large-model era rewarded exactly its specialisations. When the most important technology stories in China became, week after week, a new model, a new robotics company, a new financing round or a newly famous researcher, a publication founded in 2017 to cover precisely those things found its moment.

How to read and use QbitAI

QbitAI is best used as a fast, well-connected record of what China's AI industry is doing and saying now. Its news stories are valuable for names, figures, product details and the sequence of events in a fast-moving story, and its framing often reveals how the Chinese sector itself understands a development. Claims originating with companies, particularly training costs, benchmark comparisons, valuations and user numbers, should be attributed to their source and checked against technical reports, exchange filings or the companies' own announcements before being relied on for investment or research decisions, a caution that applies to all rapid technology news. Its think tank reports deserve separate treatment as research publications: their survey findings, such as the IDC collaboration's talent results, can be cited as data, with their methodology and date kept in view. For English-language readers encountering QbitAI through translation or syndication, it provides something difficult to obtain elsewhere, a sense of the Chinese AI world's own daily priorities, written in Beijing since 2017 by journalists for whom that world is the home beat.