The cost burden of AI is finally being passed on to ordinary consumers.

On June 25, Apple raised prices for several MacBook, iPad, and other hardware products. Around the same time, Microsoft Xbox announced that it would increase the prices of gaming consoles globally starting August 1.

The simultaneous price hikes by these two consumer electronics giants have brought a hardware inflation, previously hidden in servers, data centers, and supply chain contracts, more clearly to the forefront of the end market.

This is not an ordinary product price adjustment.

In the past, when capital markets talked about AI, the focus was more on large models, GPUs, computing power leasing, optical modules, and electricity. The stronger the performance of chip companies, the more willing the market was to believe that this technological cycle was still expanding.

However, the price increases in consumer electronics indicate that the impact is no longer confined to the cloud.

Data centers' sustained purchases of memory, SSDs, and enterprise-level storage are reshaping the global allocation of hardware resources, affecting consumer devices such as PCs, smartphones, tablets, and gaming consoles through cost transmission.

Chinese enterprise Lenovo Group also stated at its investor day that the current changes in the memory market are "unprecedented," with supply shortages set to persist and price increases potentially becoming the new market norm.

Consumers see computers becoming more expensive, while the industry sees a reshuffling of hardware pricing power.

Apple's price hike carries a strong signaling effect.

As the global consumer electronics company with the strongest supply chain bargaining power, Apple typically has the ability to buffer cost fluctuations through advance orders, inventory management, product portfolio adjustments, and supplier negotiations.

When Apple publicly attributes the price hikes to memory and storage cost pressures, it indicates that this round of price increases has exceeded the internal absorption capacity of leading end-device brands.

According to Reuters, Apple raised the price of the 512GB MacBook Air from $1,099 to $1,299, the 1TB MacBook Pro from $1,699 to $1,999, and the 128GB iPad Air from $599 to $749.

Apple also stated that it had been absorbing cost increases for consumers over the past period but has now reached a stage where it must raise prices for some products.

Microsoft Xbox's price adjustment further proves that the pressure is not limited to a single manufacturer or to computers.

Microsoft stated that due to significant increases in storage and memory costs, the Xbox 512GB model will see a $100 price hike, while the 1TB model will increase by $150. Gaming consoles are more directly impacted by sudden hardware cost increases.

The core issue worthy of attention is no longer short-term supply shortages in the traditional sense but rather how AI data centers have altered supply chain priorities.

In the past, PCs and smartphones were among the most important sources of demand for memory chips. Consumer electronics giants, with their stable shipments and large scale, held a strong position in the supply chain over the long term.

In the AI era, scenarios such as servers, enterprise-level SSDs, HBM, and high-capacity RDIMMs offer higher profit margins and certainty. Cloud vendors, AI companies, and chip giants are willing to lock in supply through long-term agreements, prompting memory manufacturers to prioritize these customers.

Previously, TrendForce forecast that traditional DRAM contract prices would rise by 58% to 63% quarter-on-quarter in the second quarter of 2026, while NAND Flash contract prices would increase by 70% to 75% quarter-on-quarter.

More critically, these price hikes are accompanied by shifts in production capacity. DRAM suppliers continue to redirect capacity toward server-related applications, while NAND capacity is increasingly allocated to enterprise-level SSDs, forcing consumer applications to contract under cost pressures.

Gartner's forecast is even more direct: by the end of 2026, the combined prices of DRAM and SSDs will rise by 130% compared to 2025, driving PC prices up by 17% and smartphone prices up by 13%. This will result in a 10.4% year-on-year decline in global PC shipments and an 8.4% drop in smartphone shipments in 2026.

In Gartner's view, low-cost PCs will be particularly hard hit, with the sub-$500 entry-level PC market potentially disappearing before 2028.

This means that hardware price hikes driven by AI will alter consumer replacement cycles, compress the space for entry-level products, and force end-device manufacturers to choose between raising prices, reducing specifications, delaying launches, or sacrificing profits.

Over the past two years, market discussions about AI have focused on who could invest more capital. Now, the focus is shifting to who will bear the increasingly high hardware costs.

Shawn Kim, head of Morgan Stanley's European and Asian technology team, refers to this phenomenon as "chipflation," or chip inflation. In his June 8 report, The High Cost of AI Memory, he mentioned that AI infrastructure construction is reshaping the structure of memory demand. AI has turned storage into one of the most scarce resources, with data centers taking an increasingly large "slice of the memory cake."

According to Morgan Stanley's calculations, by 2028, servers will account for 59% of DRAM demand, significantly higher than the 37% in 2023. Enterprise-level SSDs will account for 65% of NAND demand, up from around 18% previously.

This will create a real supply gap. By 2027, PC memory demand may face a 15% shortfall, equivalent to about 58 million PCs. Smartphone memory demand may face a 12% shortfall, equivalent to about 134 million phones.

Domestic securities firms' judgments point in the same direction. CITIC Construction Investment believes that from 2026 to 2027, HBM, DRAM, NAND, and small-capacity storage may all experience varying degrees of supply shortages. This round of memory price hikes differs from previous cycles, with both the duration and magnitude of price increases likely to far exceed expectations.

When applied to the industrial chain, these judgments create clearer stratification. Upstream memory manufacturers, enterprise-level SSDs, server memory, and semiconductor equipment sectors may continue to benefit from price increases and expansion demand. However, end-device manufacturers such as PCs, smartphones, tablets, and gaming consoles will face greater tension between rising costs and weakening demand.

Datong Securities mentioned in its TMT industry weekly report that the rise in the memory chip sector is essentially the result of the gradual spread of AI capital expenditures. The market is moving beyond GPUs and AI servers to further explore supporting links in data center construction, with memory chips being one of the more certain directions. However, it also cautions that the storage industry remains cyclical, and the sustainability of the rally ultimately depends on whether AI demand can continue to materialize and whether manufacturers can maintain an effective balance between expansion and pricing.

A cost calculation from the cloud to the end device is now officially underway.

More subtly, rising hardware costs are beginning to erode capital market sentiment.

Around the time Apple announced its price hikes, AI chip stocks in the U.S. market experienced a collective pullback. On June 23, the Philadelphia Semiconductor Index fell 7.9%, with Micron Technology dropping 13%, Nvidia falling 4.1%, and other chip stocks like Qualcomm and Marvell also under significant pressure.

Reuters reported that this pullback followed a sharp rally in AI infrastructure-related stocks, with markets now concerned about the sustainability of AI capital expenditures, valuations, and crowded funding trades.

On June 26, selling pressure on chip stocks had not fully dissipated. The Philadelphia Semiconductor Index fell 5.29%, with On Semiconductor dropping 24%, its largest single-day decline since 2020. Western Digital, Seagate Technology, SanDisk, NXP, Qualcomm, Micron Technology, and Applied Materials all saw significant declines.

Take Micron as an example. Its performance is not poor, and memory price hikes even benefit its profitability. However, the market's focus has shifted from "price hikes bringing profits" to whether "AI hardware inflation will suppress end demand and increase capital expenditure pressures."

At the same time, news of OpenAI delaying its IPO added to this sentiment.

Reuters, citing The New York Times, reported that OpenAI is considering delaying its public listing until 2027 to secure a higher valuation. Advisors have presented two options to company executives: wait until 2027 for a $1 trillion valuation or go public sooner at a lower valuation.

This news struck at the market's most sensitive point: AI company valuations are rising, but commercialization and cash flow validation still require time.

Apple's price hike, Xbox's price hike, soaring memory prices, a pullback in U.S. chip stocks, and news of OpenAI delaying its IPO may seem like unrelated stories, but they all point to AI entering a phase of "cost calculation."

This logic can also be mapped to the A-share market.

On the evening of June 26, Xiechuang Data announced plans to issue shares to specific investors to raise no more than 8 billion yuan for projects including a smart computing center, data storage expansion and upgrades, as well as working capital replenishment and bank loan repayment.

Xiechuang Data's private placement is highly representative in the current context. On one hand, it shows that computing power and storage remain directions capital is willing to support. On the other hand, it cannot be ignored that computing power is not a light-asset story.

Overseas end-device manufacturers are raising prices, while A-share computing power companies are financing for expansion. While the directions seem different, both are responding to the same pressure: as hardware becomes more expensive, companies must find ways to absorb costs.

A series of new questions arise: Can orders cover costs after hardware becomes more expensive? Can stable cash flows be generated after projects are completed? Can depreciation and financial expenses from expansion be absorbed by revenue growth?

This will determine the true differentiation within the AI hardware chain. AI is bringing about a cost redistribution from upstream to end devices.

Computer price hikes are just the beginning—a more visible signal.

From data centers to MacBooks, from enterprise-level SSDs to ordinary consumers' device replacement budgets, AI is not only transforming the cloud but also altering the cost structures of every computer, every hard drive, and every hardware company.