𝐀𝐧𝐭𝐚'𝐬 𝐀𝐈 𝐓𝐚𝐤𝐞𝐬 𝟏𝟓 𝐒𝐞𝐜𝐨𝐧𝐝𝐬. 𝐈𝐭𝐬 𝐁𝐞𝐬𝐭 𝐁𝐞𝐭 𝐓𝐨𝐨𝐤 𝐅𝐢𝐯𝐞 𝐘𝐞𝐚𝐫𝐬.

At Anta, China's largest sports group, with domestic sales twice those of Nike, a designer still draws the sketch. Its own AI model then turns it into a design draft in 15 seconds.

Of the major sports brands in China, Anta has made the most public bet on artificial intelligence. Its "AI 365" plan runs from products in 2025 to retail in 2026 and operations in 2027, and it targets an internal AI adoption rate above 50% and over 5 billion yuan of value within three years.

Its August results are company figures, but specific ones:

- Linglong, its internal AI model trained on more than 10 million sets of Anta's own design data of the past 30 years, cuts the time to extend a bestseller by 80%. As of November 2025, there were over 2.5 billion yuan of orders for designs made with it.
- A second model, Lingxi, suggests outfits to customers, and in a Tmall test store it cut returns by 10%, lifted conversion by 10% and got 20% more click-through.
- In Ops, its AI already reads social media and consumer trends to support forecasting and inventory.

The foreign brands use AI more quietly in China. adidas gets trend reports every week, and its chief executive, Bjørn Gulden, says AI now helps it predict trends and can inform silhouettes, materials and colors. Nike's most visible AI move this year, shopping inside Google's Gemini, covers the United States only.

Yet behind Anta’s biggest successes have been human decisions. None illustrates this better than FILA. In 2009, Anta bought the China rights to a brand with 50 stores and 32 million yuan of losses. A new team took it upmarket, a positioning many people thought risky, and it turned a profit only in 2014. Someone made that bet and stuck with it through five years without a profit. In 2025, FILA brought Anta 28.5 billion yuan of revenue!

Anta seems to know where the line sits. Its co-chief executive, Lai Shixian, says the group is "still at a relatively early stage" with AI. It aims to use AI for data and efficiency while people keep "creative inspiration and emotional connection."

In September, I argued that the plus in China's "AI+" is a company's own know-how ("You Bought The Model. You Did Not Buy The Capability."). Anta shows the next step: the model widens the choice, and the know-how translates into knowing which option to back.

AI multiplies the options. A person still has to choose one and own the decision.

Treat Anta as a benchmark, not a template, and run a two-list test on next season's range. Extensions of what already sells go to AI, and you measure the days saved; new bets each get a named owner who can say yes, and an agreed time to prove themselves.

#ChinaRetail #ChinaSpeed #ChinaAI+ #Sportswear #adidas

Contact

bruno.gentil@sherpaconsultingasia.com

© 2026. All rights reserved.