China's Tech Giants Are Paying AI Interns 5,000 Yuan a Day
A survey of AI hiring by the business outlet 36kr found doctoral interns at ByteDance, Tencent and Alibaba earning up to 6,000 yuan a day — more than many white-collar workers make in a month — as China's AI talent war pushes starting researcher pay past 2 million yuan a year.
A graduate student who has not yet left campus is drawing 5,000 yuan a day as an intern — roughly 700 US dollars, and more than many Chinese office workers earn in a month. If the internship converts to a full-time offer, the annual package can reach 3 million yuan. The figure opens a 36kr survey of China's AI hiring frenzy published this weekend, and it is not presented as an outlier: at the AI core businesses of ByteDance, Tencent and Alibaba, the piece reports, doctoral interns now command 5,000 to 6,000 yuan a day, while ordinary AI interns draw 500 to 1,000 plus a housing allowance.
The escalation is the story's real point. A year ago, the baseline a run-of-the-mill algorithm researcher would accept was around 1 million yuan; this year, 36kr reports, it has moved to 2 million and above. Demand-side data point the same direction: Zhilian Zhaopin, a major recruitment platform, counted 26.3 percent more companies hiring in artificial intelligence in the first half of 2026 than a year earlier, with AI job postings up 10.6 percent.
To an outside reader, the striking part is less the money than what it inverts. Internships in Chinese white-collar life are traditionally near-worthless as compensation — the word shixi (实习) carries an expectation of token pay or none, with the exchange running the other way: the intern supplies cheap labour and docility in return for a line on the résumé. Positions that pay an intern more than a mid-career office worker flip that cultural default completely, and the job boards have responded accordingly; the 36kr piece's title asks what a "1,000-yuan-a-day AI intern" has to be anxious about, a question that only makes sense once the pay has stopped being the problem.
The anxiety, going by the data in the piece, is structural. The gold rush is concentrated on a narrow band of credentials — PhDs from top programs in machine learning — while the candidate pool behind them is enormous: China is projected to produce 12.7 million university graduates in 2026. Salaries are being bid up fastest precisely where supply is scarcest, which means the boom visibly bypasses most of the graduates crowding into the field, and the winners hold seats whose value depends on technology that is still being rewritten. Hiring figures where postings grow more slowly than the companies posting them (10.6 percent against 26.3 percent) hint at firms consolidating AI teams even as they bid up individual stars.
The story ranked near the top of 36kr's board over the weekend, one of several signs that China's AI race has entered its talent-auction phase. For an international reader, it is the clearest available price signal for what the country's big platforms believe large models are about to be worth — and of what a fresh PhD in the right subfield currently costs.