What is the connection between commodity price volatility and AI-related stocks?

Since the tech stock rebound began on July 31, 2026, I’ve increasingly sensed intense daily battles between bulls and bears. This can be clearly seen from intraday charts: every time a rebound occurs, selling pressure immediately emerges to suppress it. Some of this selling pressure comes from investors who have been stuck in losses since earlier highs—now seeing a rebound, they rush to cut losses and exit. Others are short-term speculative funds that entered near the bottom and, having realized profits, now seek to lock in gains. Such trading behavior leads to progressively healthier rebound structures, as shareholding positions continuously rotate—each day’s position structure remains sound. This qualifies as a relatively comfortable rebound phase. As for those exiting, from a technical analysis perspective, they’re clearly making a mistake. Take the recent performance of the communications sector: it has steadily advanced along its moving averages—a textbook-perfect pattern. It’s now approaching the critical juncture where the 20-day and 120-day moving averages converge. Few investors are paying attention right now; sentiment is broadly pessimistic. But once the index breaks above the 20-day MA, many may suddenly catch the scent and rush in. That’s human nature: people rarely buy when stocks are cheap—only after substantial gains do they feel comfortable jumping in. Most currently believe the tech rally has peaked—a view evident both in market action and online discourse. I’ve consistently maintained my strong conviction that the AI-driven bull run is far from over. Since June, I’ve repeatedly stated this clearly: we’re experiencing a large-scale deleveraging correction. If you’re skilled enough to exit early during this correction, that’s ideal. If you inadvertently remain invested, it’s not fatal—because there remains a solid chance of eventually recovering close to previous peaks. I’ve emphasized repeatedly: rebounds lasting over two weeks, reaching near the 60-day MA, and initiating from above the 250-day MA—all these criteria have been met. I’ve also noted that tech stock rebounds inevitably align closely with gold and silver rallies, as both reflect the same underlying risk appetite. The same pattern occurred in March 2026: gold first stabilized and rebounded, followed by a major tech rally. Many investors focus solely on stocks and ignore commodity markets—hence they miss these linkages. Another recent development worth noting is copper prices beginning to rise. With potential U.S. copper tariffs looming in 2027, importers are rushing to secure supplies ahead of time—triggering a ‘preemptive import’ surge. Observe copper’s price chart: it’s rising cleanly along its moving averages. The fundamental driver behind rising copper prices is surging U.S. demand for copper—essentially, massive growth in AI infrastructure investment. Thus, rising copper prices provide independent confirmation that AI infrastructure expansion has not slowed. In commodity markets, key indicators include gold, silver, copper, aluminum, and oil.

Is the shift of capital from AI hardware to AI applications a relay handoff—or an exit?

The second category is AI for Science—targeting a small cohort of scientific researchers—with representative directions including AI-powered drug discovery and new materials R&D. Individual projects here carry exceptionally high value, rely heavily on top-tier large models, and while domestic AI-pharma candidates have already entered Phase II clinical trials, scalable profitability remains distant in the near term. The third category is mass-market multimodal entertainment—including AI-interactive video, virtual avatars, intelligent companionship, and game interaction. This represents a high-token-consumption, consumer-facing segment. Companies like miHoYo and Alibaba continue investing in both computing power and product development, offering vast long-term potential. III. Investment Framework for AI Applications: Anchored in Payment Responsibility, Value Creation, and Verifiable Performance AI application investment logic differs markedly from upstream compute hardware—and cannot simply mirror capital expenditure or capacity allocation metrics. When screening candidates, prioritize answering three core questions: Who bears the payment cost? What incremental value does the product deliver? Can ROI be quantified and measured? By this standard, marketing and general office productivity applications exhibit the clearest commercial closed-loop. Sector analysis demands deep industry specialization and on-the-ground validation of real-world product deployment—conclusions cannot be drawn quickly, making this approach better suited for medium- to long-term tracking. Moreover, China’s AI applications possess unique competitive advantages: leading domestic open-source models closely track global front-runners, and the industry—operating under constrained compute supply—has honed exceptional cost-control capabilities. Product internationalization may replicate the cost-performance advantages previously demonstrated by Chinese consumer electronics and short-video platforms. At the industrial chain level, upstream domestic compute hardware still offers medium- to long-term opportunities—covering chips (e.g., Cambricon, Hygon), compute leasing (e.g., Xiechuang Data), and switching chips (e.g., Shengke Communications). Notably, CPUs—critical for serial logic operations in AI Agents—serve as a vital bridge between compute and applications, suggesting room for increased allocation weight in portfolios. Downstream AI applications span general office productivity, cybersecurity, tax & finance, industrial automation, multimodal content, marketing, and AI-pharma—represented by firms including Kingsoft Office, Fanwei Network, Wondershare, Sangfor Technologies, TaxFriend, SUPCON, Heshi Information, Maifushi, and Kingdee International. P.S.: Cyclical sectors see profit-taking; optical communication leads, validating compute sector strength Today’s market displayed clear capital reallocation: cyclical sectors—including oil & gas and coal—faced pressure, with sustained outflows from resource commodities. Meanwhile, the communications equipment sector surged 4.77%—topping all sectors. Yonding Co., Ltd. hit its daily limit with ¥8.165 billion in turnover; Tongyu Communications posted five gains in eight days; Juguang Technology achieved a 20%涨停; Zhongji旭创 closed at ¥921, with ¥28.408 billion in turnover; Tianfu Communications rose sharply in tandem. On the earnings front, Foxconn Industrial Internet disclosed its semi-annual report on August 11: AI server business drove net profit attributable to shareholders to ¥23.74 billion—a 96% YoY increase—directly confirming continued strength across the compute supply chain. From a funding perspective, Dongshan Precision, Advanced Micro-Fabrication Equipment (AMEC), and Jingwang Electronics saw substantial net inflows from major institutional players; margin financing and Shenzhen-Hong Kong Stock Connect funds simultaneously bought aggressively—forming a coordinated push into tech hardware. Copper prices remain elevated, enabling smooth cost pass-through for upstream PCB material suppliers—pushing profits further upstream. Judging from trading volume composition, this rebound has been led primarily by mid- and small-cap PCB and fiber-optic cable stocks—indicating institutional rebalancing from overvalued large-cap leaders toward downstream and midstream segments, rather than mere technical recovery.

Copper Steady as AI-Driven Tech Rally Holds Near Record

Oct 6, 2026

Copper prices held steady on the LME supported by AI data center demand expectations. Markets remain cautious ahead of China's reopening following recent tariff-driven supply shifts.

bloomberg

Copper Steady as AI-Driven Tech Rally Holds Near Record

FAQ