CONNECT WITH US
Jul 21
WAIC 2026 elevates the 'token factory' — why is China chasing cheap tokens over bigger models?

At the 2026 World Artificial Intelligence Conference (WAIC) in Shanghai, held July 17-20 across "three districts and four venues" with more than 1,100 exhibitors, one phrase recurred across vendor booths and keynotes: the "token factory".

ASEAN's AI boom is taking on a dual role as both a "no-man's land" and an international hub from Singapore's perspective, while investors from the Middle East and China are also pushing into the region. According to industry sources, ASEAN's AI industry will be driven not only by local demand, but also by China's market.

The forthcoming AI dialogue between Washington and Beijing is less a sign of easing tensions than of a changing rivalry. Both governments are tightening their grip on frontier AI, extending competition beyond model development into the rules governing the technology.

The US and China are preparing to hold their first official artificial intelligence (AI) dialogue under President Donald Trump, a notable diplomatic milestone after years of escalating technology tensions. Yet the significance of the meeting lies less in whether it produces tangible agreements than in what it reveals about how both countries increasingly view AI: not as another fast-growing industry, but as strategic infrastructure requiring tighter state oversight and national control, according to Reuters.

Samsung Electronics executive chairman Lee Jae-yong, SK Group chairman Chey Tae-won and Naver founder and board chairman Lee Hae-jin are reportedly preparing to meet Nvidia CEO Jensen Huang near Silicon Valley, with a roundtable tentatively planned for July 24.

LG Electronics plans to stop production at its Huizhou plant in China after October, ending manufacturing at the group's first production subsidiary in the country after more than three decades.

Wistron officially opened its Dallas, Texas plant on July 21, becoming the company's first mass-production facility in the US and the first US mass-production plant among Nvidia's ODM partners. Already producing GB300 baseboards at volume, the plant will expand to Vera Rubin architecture products in the future.

The AI industry may be riding a wave of momentum, but two clouds have recently unsettled the outlook: growing public opposition to data center construction across the United States, and a dramatic leap in the rankings by Chinese-developed model Kimi K3, which has renewed fears of AI overinvestment. NVIDIA CEO Jensen Huang addressed both concerns directly on July 21.

Nvidia's newly detailed Vera Rubin NVL72 is being pitched less as a faster chip and more as a cheaper unit of AI output. This shift matters most for the power-constrained data centers now defining the ceiling of the industry's growth. In its July 21 announcement, Nvidia said the rack-scale system delivers roughly 10 times more tokens per megawatt than its current Grace Blackwell NVL72, and about one-tenth the cost per million tokens of the GB200 NVL72. This framing positions electricity, not silicon, as the scarce resource.

Microsoft's decision to pour billions of dollars into Mistral's European compute base is less about adding another model to Azure than about answering a question that has grown loud across the region's boardrooms and ministries: can enterprises run frontier AI without ceding control of their data to US-operated infrastructure?

Chinese AI chip developer Enflame said at the World Artificial Intelligence Conference (WAIC) 2026 that it had jointly showcased an AI chip sample with Shanghai Xianfeng Technology using glass-based Chip-on-Panel-on-Substrate (CoPoS) panel-level advanced packaging.

A funding-and-listing push by Beijing-based Moonshot AI signals that China's frontier-model race is spilling over into the capital markets, with implications for global AI valuations, the open-weight debate, and Nvidia's China-exposed customers.