China wants to be an AI superpower. It also wants to keep nearly a billion workers employed. A growing tension between Beijing’s employment-first mandate and its technological ambitions is forcing companies to pump the brakes on automation, even when the math clearly favors machines over humans.

The court ruling that set the tone

In May 2026, a Chinese court issued a ruling that sent a clear signal to the corporate world: companies cannot dismiss workers simply because AI can do their jobs.

The ruling didn’t emerge in a vacuum. It followed directives from China’s State Council in August 2025 calling for stronger employment risk assessments tied to AI deployment. Before a company rolls out automation, it needs to prove it won’t gut payrolls in the process.

Youth unemployment in China sits at an estimated 16-17%, and roughly 12.7 million graduates are expected to flood the job market in summer 2026.

When robots do 90% of the work but nobody gets fired

JD.com offers perhaps the most vivid illustration of the contradiction at play. The e-commerce giant, which employs around 900,000 people, has pledged to safeguard jobs across its workforce. This commitment exists alongside the fact that robotics have already replaced 90% of workers at one of JD.com’s sorting centers.

Companies that can’t restructure openly may simply slow-walk innovation instead. Rather than deploying AI at full speed and retraining displaced workers, they keep humans in roles that machines could handle more efficiently.

The numbers behind the squeeze

According to estimates from Citi, approximately 9.6% of Chinese jobs, roughly 70 million roles, are highly exposed to AI displacement.

What this means for investors

For anyone with exposure to Chinese tech firms or broader AI plays in the region, this policy dynamic deserves close attention. Companies operating under Beijing’s employment framework face fundamentally different constraints than their Western peers. Innovation cycles may stretch longer. Restructuring that would take quarters in the US could take years in China, if it happens at all.

The risk isn’t that Chinese AI development stops. The risk is that the gap between AI capability and AI deployment widens. A country can have world-class models and still underperform on productivity if policy prevents those models from being used to their full potential.

With 70 million jobs at high risk of displacement and 12.7 million graduates entering the market in summer 2026, the current equilibrium feels inherently unstable. Something will have to bend, either the commitment to job preservation or the ambition to lead the world in AI.