Zoom out and two more threads landed in the same stretch: in May, 16 American CEOs flew to Beijing with President Trump, voting with their capital; around the same time, McKinsey published a report on how the Asia-Pacific computing landscape is being reshaped, with Australia vying to become a computing hub. Taken together, the four events point the same way: the next phase of globalization is neither simple "decoupling" nor "co-governance," but a multipolar, selective competition. The cards China's going-global companies hold are hidden in this structure.

One. A single company's lesson: Nike treated DTC as an on/off switch

Nike's DTC (direct-to-consumer) transition ran for thirteen years — launched in 2017, exited Amazon in 2019, and cut nine retailers in 2020. By 2021, DTC revenue reached $12.4 billion, roughly 35% of total. Then the trouble began.

The dropped Foot Locker turned its shelves over to On and Hoka, whose revenues each climbed to about $2 billion. Nike itself leaned on retro reissues of designs from 40 years ago; product innovation lagged. BNP Paribas analyst Laurent Vasilescu put it bluntly: "Nike's problem isn't distributors, it's product."

By 2026, Nike went further in China — cutting thousands of online distributors, with Topsports shares briefly sliding over 20%. Nike treated DTC as a switch: turn off distributors, turn on official stores, and you're done. But DTC is not a switch; it's a capability system.

Whether a company fits a given path depends on whether product strength, brand momentum, and capital reserves are all in place. Nike in China logged multiple quarters of negative growth, leaned on old designs, and cut distributors without figuring out how to fill the gap — with all three missing, DTC became a tool that accelerated its weak spots.

By contrast, brands "born DTC" like Anker and Roborock put product first and channel second, pricing on technology premium rather than logo premium. Anta and Lululemon run "omnichannel symbiosis," with direct retail, distributors, and e-commerce each doing their job — more resilient. No path is inherently right or wrong; the key is the sequence: product strength → pricing power → channel strength. Nike's mistake was moving the channel first.

Two. A signal from a group of companies: 16 CEOs on one plane

On May 14, President Trump visited China for the first time in nine years, accompanied by a CEO roster released by the White House — Apple, Tesla, Qualcomm, Micron, Meta, Boeing, GE Aerospace, BlackRock, Blackstone, Goldman Sachs, Citigroup, Mastercard, Visa, Cargill. Technology and finance in equal weight.

Two details stand out. NVIDIA's Jensen Huang was not on the original list and joined just before departure; the image of him boarding with a backpack was captured worldwide. JPMorgan CEO Jamie Dimon missed the trip at the last moment. The absentees and the latecomer are themselves a signal.

Equally telling is who was not there: Google, Microsoft, Amazon, and other pure internet and cloud giants were absent. The traveling group clustered in hard-tech manufacturing that needs China's supply chain and financial capital betting on the Chinese market for the long term.

Then look at hard numbers: by the U.S. Department of Commerce's count, China–U.S. bilateral trade still exceeded $600 billion in 2025. In the same period, China's electric vehicle exports overtook fuel-vehicle exports for the first time. While Washington talks "decoupling," business is voting with capital and supply chains. It also shows that the China–U.S. divide is not about "whether to cooperate" but "on which layers to cooperate, and on which layers to keep distance."

Three. A regional layout: the Asia-Pacific computing map is being reshaped

In April 2026, McKinsey published "Australia's AI moment." Its core judgment: global computing demand will grow at least 3.5× by 2030; traditional data-center markets in the U.S., Europe, and Singapore are hitting bottlenecks in power, land, and approval timelines, so demand is spilling over to nearby regions.

Who is absorbing that spillover? Malaysia's Johor region drew over $20 billion of North American hyperscaler investment in the first ten months of 2024; Thailand approved about $10 billion of data-center investment in H1 2025; India and Indonesia are following. Australia wants to spend A$190 billion in new capital to lift its computing capacity from 1.5GW to 5.0GW — a public, quantified ambition.

This map cuts two ways for China. On the training side, which demands extreme-density compute and chases cost advantages, export controls may raise China's "relative cost." On the inference side, which demands low latency and proximity to users, China's 1.4 billion people mean local inference demand is extremely strong — an advantage no other country can easily replicate. Saying "the computing landscape decides the AI competition" is not much of a stretch.

Four. The Chengdu statement: open the door, or shut it?

Put the three threads back together, and the Chengdu statement's answer is "open the door" — but with careful posture.

The ministerial statement uses words like "encourage," "recognize," "note" — in diplomatic language, "soft law" with no binding force. But writing "open source" into an APEC ministerial statement for the first time signals momentum: when unilateral sanctions try to restrict the flow of open-source AI models, 21 economies sitting together saying "we choose openness" is itself a statement.

This also hands China's going-global companies a card. In the past, Chinese globalization rested on the World Factory's cost advantage; it is now shifting toward "technology export + participation in rule-making" — the Chengdu statement, RCEP, and Belt and Road infrastructure are all part of that hand.

For companies, this card can only lower the cost of setting sail, not the risk of the voyage. In Forbes China's 2026 DTC cross-border influence list, over 60% of the companies used teams of under 300 people to reach billion-scale overseas revenue — but those are the top performers. The far larger base of small and medium firms faces tariffs, return rates, and platform fees that land hard on the ledger.

So the card is not a "winning hand" but a position: in multipolar competition, the opportunity for Chinese companies lies not in "choosing a side," but in holding a playable card in every direction.

References

  1. 2026 APEC Digital and AI Ministerial Statement (Chengdu Statement) — APEC, 2026-07-23
  2. Nike to cut off thousands of online distributors in China — CNBC, 2026-07-21
  3. Nike moves to cut 'thousands' of online retailers in China — Reuters, 2026-07-22
  4. Nike FY2026 Q4/FY earnings: Revenue flat, Nike Direct drops 6% — CNBC, 2026-06-30
  5. Nike Is Cutting Off More Than 1,000 Online Third-Party Sellers in China — WSJ, 2026-07-22
  6. Nike reverses DTC strategy, re-embraces wholesale partners — Modern Retail, 2024
  7. Xi Jinping meets Trump and US business delegation — Xinhua, 2026-05-14
  8. White House releases CEO delegation list — CNR (央广网), 2026-05
  9. H200 chip exports to China approved — Reuters, 2026-05
  10. Australia's AI moment: Building Asia–Pacific's compute hub — McKinsey, 2026-04
  11. U.S., other nations back open-source AI with 'strong security' at China summit — CNBC, 2026-07-24
  12. Anta 2025 results: revenue RMB 80.219 billion, surpassing Nike China — Xinhua/TMTPost, 2026
  13. Roborock Becomes the World's No. 1 Smart Cleaning Robot Brand — PR Newswire/IDC, 2026-03-12
  14. Forbes China 2026 DTC Cross-Border Influence Brand List — Forbes China, 2026-07

💡 What did this article inspire for you?

humanaifit studies how humans and AI can genuinely work together. If you face real questions on enterprise AI adoption, human-AI collaboration, or global compliance, join our discussion.

🔗 Search for the "AI Era Survival Handbook" Knowledge Planet, ¥199/year — every deep article comes with tool templates and direct contact with the author.