According to informed market sources, the E Fund Asian Semiconductor ETF (03486) has surged by more than 2%. As of the latest update, it is trading up 2.06% at HK$20.3 per share.
On the news front, on September 9, OpenAI announced that its internal AI system has successfully resolved the existence and smoothness issues of the Navier-Stokes equations—one of the seven "Millennium Prize Problems" that had remained unsolved for nearly nine decades. Official disclosures reveal that the system utilized approximately 10,000 concurrent agents, taking roughly 88 hours from launch to find a solution, with an additional 17 hours dedicated to formal verification in the Lean language. Concurrently, OpenAI also unveiled ChatGPT Images 2.5, which brings upgrades in speed, fidelity, and editing consistency for image generation.
The significance of this breakthrough lies not in the mathematics itself, but in the approach and its broader implications. First, AI capabilities are shifting from "conversational generation" to "scientific discovery." Second, the narrative surrounding hardware demand has been further solidified. The pre-training phase for GPT-6 consumed more than 100,000 Blackwell GPUs, and long-horizon tasks like computer-use could entail token consumption that is ten times the scale of AI coding. Third, in the week since GPT-6 has been in circulation, its demonstrated abilities have continued to generate excitement, with the market closely watching whether it can replicate the commercial success loop seen in coding.
In summary, the AI industry is currently in a positive feedback cycle of "capability acceleration—narrative diffusion—rising hardware expectations." Scientific breakthroughs and multimodal upgrades are opening new application scenarios for AI, and behind every leap in capability lies more intensive computing power usage, which bodes well for the valuation and prosperity outlook of the global AI computing and semiconductor supply chain.
When it comes to asset allocation, there are two clear investment tracks to consider. The expansion of AI inference demand directly fuels prosperity in memory and advanced process nodes. For flexible exposure to the hardware side, investors might look at the E Fund Asian Semiconductor ETF (03486), which focuses on the full Asian semiconductor chain, including key players like SK Hynix and TSMC. For a foundational tool to capture the "AI capability acceleration plus computing demand expansion" theme, the E Fund AI ETF (03489) is worth noting. This ETF covers leading US computing stocks as well as Hong Kong-listed hard tech and application-layer companies, with NVIDIA holding a top weight of 8%.