News · Technology

GPU Prices and Battery Supply Test Buyer Choice

By AWEI · AI-compiled · Published 2026-09-16 · 2 sources · 36kr.jp, dtf.ru

Reported GPU price rises and concentrated battery shipments raise different questions about buyer choice, supply and dependence.

For mainland China graphics-card buyers, the immediate stake is the price of an upgrade. DTF reports planned Asus RTX 5070 increases of 150–300 yuan and an implemented Gigabyte RTX 5060 Ti 16GB increase of 200 yuan. Separately, 36Kr Japan cites TrendForce estimates of more than 460 GWh in first-half 2026 energy-storage cell shipments, with roughly 84% supplied by ten Chinese companies. These developments raise a shared question about buyer choice, without establishing a shared shortage.

Production is not the same as access

The distinction matters because expanding an industry does not necessarily expand every customer's options. DTF attributes to 3DCenter an account in which Nvidia reallocates GB20x Blackwell chips toward professional products rather than cutting 4-nanometer production. If accurate, that would make allocation between customers central to gaming-card availability. A factory can maintain output while the mix reaching one group changes. Buyers would then depend partly on supplier priorities, rather than simply on how many chips can be manufactured.

That explanation remains contested within the reporting. Memory costs or distributor pricing could also raise GPU prices, including without a substantial allocation shift. Reported increases affecting AMD products make the input-cost explanation worth examining, but do not establish it. Chip production, professional-product allocation and finished gaming-card shipments are separate measurements. Treating any one as a substitute for the others would make a precise price report carry more causal authority than its evidence supports.

Battery scale poses a different question

The battery story concerns concentrated growth. According to the TrendForce findings reported by 36Kr Japan, the top three suppliers accounted for about 44% of shipments, and CATL shipped more than twice the runner-up's volume. Those shares describe where cells came from; they do not measure pricing power, profitability or installed storage capacity. Dependence would become more consequential if customers could not readily qualify another supplier. Without evidence about substitution, a concentration figure identifies a question rather than answers it.

There is also a credible efficiency interpretation. Major suppliers may win business through cost, quality or dependable delivery, rather than by restricting competition. Concentration could therefore accompany benefits for buyers, even while reducing the number of practical alternatives. The residential segment cautions against a single account of the industry: 36Kr Japan reports a different leading group, comprising REPT, Great Power and EVE Energy. Procurement options may vary by application, so the overall ranking cannot describe every customer's negotiating position.

The value and cost of alternatives

The World Bank's Commodity Markets Outlook April 2026 2027, published in April 2026, offers a useful diversification lens. Its policy synthesis, informed by commodity-market analysis and econometric findings, addresses importing and exporting economies, particularly developing economies, across energy, fertilizer, LNG and critical-mineral supply chains. It considers near-term contingencies and longer-term structural alternatives. It does not measure substitution among finished GPUs or battery cells. Applied here as a question, its framework asks whether a nominal alternative is usable when a preferred supplier changes terms.

Diversification is not automatically the cheapest answer. Qualifying additional suppliers could add expense and complexity, while common input pressures could affect several suppliers together. The distributional issue is who can absorb those costs: buyers able to maintain alternatives may have more room to negotiate than those tied to one product or qualification process. That is a conditional implication, not an observed difference in these reports. Neither account establishes which buyers have substitutes or what switching would actually cost them.

The next evidence should test these explanations. Stable chip production alongside falling gaming shipments and rising professional allocation would strengthen the allocation account; prices tracking memory costs across vendors would favor the input-cost account. Battery buyers qualifying alternatives without major delays would weaken the dependency argument. TrendForce's roughly 1,012 GWh full-year shipment figure remains a forecast, and global GPU price effects remain prospective. Greater scale can support choice, but these reports leave the practical availability of alternatives unresolved.

All newsNews archive