China’s drive for AI self-sufficiency rewarding firms with global focus

China’s push for global supremacy in artificial intelligence technology is increasingly benefiting companies that supply the rest of the world, with investors favoring such stocks as fierce competition hurts their domestically focused peers.

A Bloomberg gauge tracking 30 Chinese technology stocks with the biggest overseas revenue exposure has delivered a return of 36% this year, versus 9% for those more dependent on local sales. A separate measure of the export-oriented firms’ outperformance over the other group is set for the strongest-ever reading this year.

The contrasting fortunes partly result from Beijing’s ambition to build its own AI ecosystem. The self-sufficiency campaign has led to intense local competition and a vicious price war that has eroded profit margins on chips to robotics. Meanwhile, export-oriented Chinese companies have thrived on surging global demand for AI infrastructure such as data centers.

The future of China’s AI industry is back in the spotlight this week ahead of a closely watched summit between U.S. President Donald Trump and his Chinese counterpart Xi Jinping. All eyes are on whether the two countries can put guardrails around the disruptive technology amid a heated race for global dominance.

“Domestic competition is tough no matter what industry. That’s the major concern,” said Elinor Leung, managing director of Asia telecom and internet research at CLSA. “If you can sell internationally, the margin is much higher.”

Among the big winners of China’s AI export boom are optical component makers Zhongji Innolight and Eoptolink Technology, both of which derive more than 90% of their revenues overseas. The two stocks have each rallied about 50% this year.

Fueling the gains are a strong outlook for global AI spending. Nvidia Corp. CEO Jensen Huang recently said that he expects the company to sell twice as many chips in the coming year, while the initial success of Meta Platforms’ new AI agent offers further evidence that demand for semiconductors and computing power will likely stay robust.

Many emerging-market funds remain underexposed to China’s domestically focused sector and favor exporters leveraged to the U.S. AI capital expenditure cycle, Bank of America strategists wrote in a note last week.

Chinese AI firms with an inward focus, known as the localization trade, have lost steam in recent years after the industry fell victim to cut-throat domestic competition that also engulfed other sectors from electric cars to food delivery. In order to gain market share, companies have rushed to roll out new, cheaply priced products including AI models and apps at the expense of sustainable profitability.

Shares of Moore Threads Technology, a leading Chinese AI designer that generates nearly all of its sales locally, are down about 25% this year. Those of AI tools developer SenseTime Group, with over 90% of its sales derived locally, have fallen over 40% in the same period.

Meanwhile, social media firm Kuaishou Technology, which also runs an AI video business and generates less than 5% of revenue overseas, has slumped about 51%.

To be sure, some observers still consider the China AI localization trade valid at a time of unabated geopolitical uncertainties that may hurt exports.

Morgan Stanley’s analysts wrote in a note last week that AI sovereignty and semiconductor localization remain structural opportunities for Chinese equities regardless of the outcome of the upcoming summit, while AI infrastructure, biopharmaceuticals and other cross-border sectors are most vulnerable to a reescalation of geopolitical tensions.

Still, others say the race-to-the-bottom style of competition, known in China as involution, is far from over and will continue to plague domestically oriented AI firms.

“In some industries, domestic demand is already well served, or there may even be excess supply,” said Matty Zhao, co-head of China equity research at Bank of America. “Against that backdrop, the ability to export or grow internationally can be an important differentiator.”