A June 2026 layoff wave at Ctrip and Meituan is the first Chinese tech cycle to cite AI productivity as the stated HR rationale for cuts across performance bands, a Deep Dive column report by Chinese tech outlet 36Kr finds.
Ctrip's main app used to ship a new version every two weeks, with engineering days starting at 10:30 a.m. and ending before 7 p.m. After the company's June 2026 restructuring, the cycle runs weekly and engineers routinely work past 10:30 p.m. The shift was sold internally as an "all in AI" productivity mandate. It also framed the layoffs as the first wave of Chinese tech reductions in recent memory to cut high performers and senior engineers alongside junior staff, according to 36Kr's Deep Dive column (深氪).
The investigative piece, published in late May and updated through June, traces dismissals at Ctrip (携程) and Meituan (美团) through two anonymized accounts: a Ctrip backend engineer on the hotel-commercialization team with one year of tenure and a ¥25,000 monthly salary, and a senior Meituan engineer hired from ByteDance whose entire product pod, the small team assigned to a specific feature or service, was effectively disbanded. The names in the piece, rendered here as "Lin Yue" and "Cang Shu," follow 36Kr's anonymization convention. No official layoff tallies have been released by either company; the numbers cited in the piece rest on 36Kr's tallying and on-the-record employee accounts.
What makes the cycle different from earlier Chinese tech layoff rounds is not the volume but the mechanism. Prior waves cut redundant headcount or under-performers. Senior "P-rank" engineers, known internally as 高P and the rungs above senior staff on China's tech-firm career ladder, and "exceeds expectations" reviewers were effectively insulated. In this cycle, that insulation dissolved. The stated pretext, on paper and in HR rationale, is AI productivity.
The internal reframing has a specific shape. 36Kr describes "all in AI" mandates that bundle token-usage competitions, mandatory AI training, and hidden KPIs across every product team. Teams whose output cannot be visibly tied to AI usage get flagged. Flagged pods are cut first. In one adjacent Ctrip team cited in the piece, two engineers who had received "exceeds expectations" reviews were laid off in the same batch.
The mechanism also explains a counter-intuitive choice 36Kr documents: junior engineers are often let go first, even though they cost less to terminate. The reason is the business-context gap. Junior staff can be replaced, but only engineers who understand the product, the customer, and the failure modes of the codebase can direct AI tools toward useful output. Without that context, AI assistance turns into a productivity tax rather than a productivity gain. Junior engineers without the context become the most expensive line on the team.
The tempo shift is not incidental. Ctrip's old operating rhythm included 10:30 a.m. starts, two-hour lunches, and bi-weekly app releases. That combination earned the company a reputation as an "internet retiree factory" (互联网养老院) inside the industry. The same rhythm is now the explicit target of the new mandate. Weekly shipping and late-night finishes are framed as the AI-augmented norm. The layoffs are framed as the natural consequence of fewer humans doing more work, with AI filling the gap.
The pattern rhymes with US data, but the mechanism is different. A Stanford Digital Economy Lab working paper, also hosted by the Stanford Institute for Economic Policy Research (SIEPR), finds roughly a 20% drop in employment for software developers aged 22 to 25 between the late-2022 peak and September 2025, treating the youngest cohort as a "canary" signal for the profession. The China wave runs on a different clock and a different rationale: not a labor-market statistic but an internal HR reclassification of who counts as productive when AI tools are present.
If the all-in-AI mandate spreads beyond Ctrip and Meituan, the pattern 36Kr describes becomes the new normal for Chinese tech: junior engineers without business context get cut first, and the high-P and high-performer shield no longer protects anyone.