The US and China are moving toward different models for governing increasingly capable artificial intelligence, even as the US proposes a limited safety mechanism with China ahead of President Donald Trump and Chinese President Xi Jinping's September 24 summit in Washington. AI is expected to be among the issues discussed at the meeting.
In a 46-minute interview on CBS, Nvidia CEO Jensen Huang pushed back against apocalyptic warnings regarding artificial intelligence (AI). He urged the industry to focus on rigorous engineering and safe product delivery rather than apocalyptic fearmongering.
China's BeiDou satellite navigation industry continues to expand rapidly, with applications spreading from chips and modules to end devices. Emerging technologies including artificial intelligence are increasingly being integrated with BeiDou, while the industry is shifting from hardware sales towards data, platforms, and services. Overseas markets are also emerging as a key source of future growth.
China's internet regulator is reportedly investigating DeepSeek and Moonshot AI over allegations that the two companies routed millions of user interactions, including potentially sensitive government and corporate information, to Anthropic's Claude models without customers' knowledge.
Gigabyte's AI server shipments continue to accelerate, lifting revenue and profit while prompting capacity expansion. For the first half of 2026, revenue reached NT$249.272 billion (US$7.9 billion), up 48.4% year-over-year; net profit after tax was NT$11.875 billion (US$373.3 million), up 92.6%; and earnings per share (EPS) hit NT$17.73, all record highs for the same period.
Meta's new personal AI agent, Muse, is drawing strong early consumer interest while a test involving human contractors highlights the challenges of making autonomous AI reliable and safe.
Nvidia's effort to make its autonomous-vehicle stack the default foundation for heavy freight gained a partner with paying customers on September 21, while Einride gained the compute platform it needs to expand autonomous operations across its electric freight network — a transition the Swedish company must fund through a period of heavy cash burn and a going-concern warning.
US President Donald Trump used his September 22 address to the United Nations General Assembly to reject international efforts to control artificial intelligence, argue that the US must maintain its lead over China, and urge officials to call the technology "superintelligence." The remarks sharpen a policy divide over whether rapid development can be reconciled with safeguards—and whether Washington and Beijing can cooperate on risks even as they compete.
Taiwan's export orders topped US$100 billion in August 2026, reaching US$102.96 billion, up US$5.02 billion from the previous month, according to statistics released by Taiwan's Ministry of Economic Affairs (MOEA) on September 22. Export orders for September are forecast to be between US$106 billion and US$109.1 billion, representing year-over-year growth of 50.7% to 55.1%.
China's leading AI developers are scaling revenue rapidly, but the gap with their US rivals remains substantial — and the cost of keeping pace is rising even faster. Rhodium Group estimates that seven major Chinese AI companies had combined annualized recurring revenue (ARR) of about US$10.7 billion, based on estimates dated between March and August 2026, roughly one-tenth of the more than US$100 billion attributed to OpenAI and Anthropic combined.
Meta's new AI agent, Muse, has leapt to the top of app charts roughly two weeks after its release. Its popularity has helped push up chip stocks by focusing attention on the importance of CPUs for agentic AI.
DeepSeek CEO Liang Wenfeng said during a call with investors that he aims to rely on Huawei chips to train the startup's models in the future, and that a batch of training chips from Huawei are set to begin arriving at his company as early as the fourth quarter of 2026.
Chinese labs have spent the year arguing that reinforcement learning, not pre-training, holds the next gains. Xiaomi has now put a price on that argument. The company released and open-sourced its MiMo-V2.6 series on September 22, after livestreaming the reinforcement learning post-training run that produced it — publishing step counts, token consumption, and a running cost meter on a public page as the models trained.
As artificial intelligence advances toward greater autonomy, balancing capabilities with safety has become a central challenge for industry leaders. Emerging model behaviors and high-profile security incidents have prompted a continuous evolution of safety initiatives, internal engineering adjustments, and proposed joint testing efforts between major AI developers.
