The Cost Sheet They Can’t See 

When the Huawei Pura 90 was launched in April, the company’s head of consumer business group, Yu Chengdong, in a highly unusual move, publicly revealed the pressure-inducing figures – the unit cost of this flagship model was CNY 1,200 to CNY 1,500 higher than that of the previous generation. Huawei chose to absorb this cost internally rather than pass it on to consumers, at least for now. Yu also noted that maintaining this line in the second half of the year would be even more challenging. We see it the same way. No matter how much inventory Huawei stockpiles, given its current consumption rate, it will inevitably need to purchase new components, and new components mean new prices. 

OnePlus China President Li Jie was equally blunt. At the launch event, he explained why the new product was priced this way – memory costs have risen by over 400%. Although the presentation highlighted in red text how many specifications of the OnePlus 15T have been upgraded compared to the OnePlus 15, the product’s price falls in a higher tier, at the conventional high-end threshold in China of over CNY 4,000. 

Memory is the most obvious driver, but it’s not the only component seeing price hikes. SoC costs have risen due to the use of newer chips. Display specifications are also improving, as brands compete on screen quality and need to use better panels. What sets this current cost-increase cycle apart is that memory prices are rising faster than anything else, which is quite ironic. It hasn’t been long since AI, starting with ChatGPT, began to amaze people with its capabilities, and now AI systems like OpenClaw and Claude Code are capable of assisting with complex tasks. Yet, in that short time, AI has disrupted the market for the device that provides the most global access to it – the smartphone. 

Each AI server requires far more DRAM than a single smartphone. As suppliers prioritize high-bandwidth memory for data centers, the very same foundries that once scrambled for smartphone orders are now operating on a completely different demand curve, with capacity nearly maxed out. Consumer smartphones are at the back of the line, and this queue isn’t going to clear up anytime soon. No one had modeled a scenario where AI infrastructure demands would siphon components away from devices originally designed to run AI. 

Estimated Smartphone Memory Cost Share by Price Segment, Q1 2025 – Q2 2026

Sources: Counterpoint Research BOM Service, Counterpoint Research Memory Price Tracker

A Clear Pattern of Price Increases 

Price adjustments by major Chinese smartphone brands began to surface in March. 

OPPO took action on March 16, with vivo following suit two days later. Xiaomi held out longer on paper, but it too adjusted prices for select Redmi models about a month later. Xiaomi President Lu Weibing publicly stated that the price increases were “smaller than those of peers”, effectively confirming the price hikes had already taken place. Price hikes for mid-range models ranged from CNY 200 to CNY 500, while flagship models saw even larger increases. The Find X9 Ultra launched at around CNY 1,500 higher than its predecessor. 

However, the Huawei Pura 90 is priced at CNY 4,699, essentially on a par with the Pura 80. 

Huawei Best Case Study for This Stress Test 

Huawei is the clearest example of this stress test because it had more room from the start to hold the line. 

The Kirin chips are developed by Huawei in-house. Recently, we have seen that overall supply has normalized, whether for high-end products or the newly released Enjoy 90 series, and this should begin to translate into some price advantages. Since its memory is sourced from domestic suppliers, Huawei’s cost dynamics differ from those of the international spot market. According to the company’s public statements, the domestic content of its supply chain is now close to 90%. Much of Huawei’s substantial past R&D investment involved joint development with supply chain partners, and these strategic partnerships are now paying dividends, allowing the company to secure more reliable memory supplies at better prices. The government has also provided short-term subsidies for such initiatives.  

According to a Chinese proverb, “the fragrance of the plum blossom comes from the bitter cold”. Before its resurgence, Huawei endured many hardships. But this story isn’t exclusive to China. Apple also spent a decade doing the same thing – switching from Samsung and Intel chips to the A and M series, and then spending years integrating its own baseband chips. The logic behind these stories is the same: by securing technological control, and even production control, over the key components that define the product, you are no longer at the mercy of the supply chain, while continuously requiring suppliers to provide components for which you already hold key patents. In other words, you maintain firm control over your own cost structure. Apple’s vertical integration runs so deep that even for memory, which it doesn’t manufacture itself, it secures supply in advance through large-scale prepaid contracts, ensuring favorable prices before market cycles shift. Apple was still affected by this round of supply chain disruptions, but it has more tools at its disposal, including its staggering cash flow and hardware profit margins that have led the industry for years. It used to be often said that while global smartphone OEMs essentially worked for the top few key component suppliers, Apple was the one reaping the profits. 

In this regard, Huawei has gone deeper than Apple in some areas, perhaps even further, while in others, it is still in the early stages. The underlying logic shared by both companies is this – supply chain depth is not a cost center, but an option for future competitive flexibility. 

This does not mean Huawei is immune to component inflation. Yu Chengdong has stated that cost pressures are real, and the second half of the year will bring a similar “winter is coming” moment. But compared to brands that base their product roadmaps on the assumption that “the global component market will remain accessible and predictable”, Huawei has more options when it comes to responding. We believe Xiaomi’s rapid growth is closely tied to its robust supply chain management and philosophy of symbiotic relationships within its ecosystem. However, Xiaomi’s current situation is not optimistic. Although it has adopted Apple’s “multi-supplier” system, it has overlooked how growth in another sector has turned the entire segment into a fragile, weak link. 

End of Budget Smartphone Economy 

After a period of internal and external deliberation, Meizu has essentially halted its smartphone business this quarter. The official reason given is that memory costs have risen by 150%, completely undermining the business logic of its target market segment. Setting aside the fact that Meizu was already a marginal player, the signal it sends extends far beyond a single brand. 

In the price segment below CNY 1,500, the entry-level specifications – storage, DRAM and basic configurations – that were standard just two years ago now come with significantly higher procurement costs. The number of new products launching in this price segment has noticeably decreased. The starting configurations of some surviving models have already been upgraded. Brands that once used CNY 999 or 1,299 as their flagship price points now launch new iterations starting at CNY 1,499 or higher, with gross profit margins shrinking at every step. Moreover, manufacturers are pressuring suppliers to lower prices, stating, “I am positioning this product as a budget option; at this cost, I am already not making a profit. If you can supply it, great; if not, we will cancel the project.” 

The ultra-thin smartphone category ran the same experiment at the beginning of the year. Despite massive marketing investments, initial sales volume was only a fraction of what was predicted. Xiaomi shelved its follow-up plans. Given current component costs, the market simply cannot support the price point required to build this product. 

This is not to say that budget smartphones will disappear. The Chinese market is massive and highly competitive; it won’t come to that. But the old formula for profitability – relying on manufacturing-side mass production and cost control – is crumbling. That formula, driven by falling component costs, a mature global supply chain and aggressive channel economics, has been effective for a decade. No one anticipated that AI would alter the allocation logic of memory wafer fabs, and economies of scale would no longer subsidize entry-level devices as they once did. According to the latest Counterpoint ODM report, ODMs accounted for more than 40% of global smartphone shipments in H2 2025 and were primarily impacted by memory price hikes – they are the makers of <$200 smartphones for Chinese OEMs and Samsung. ODM shipments are forecast to decline by more than 20% in 2026. 

In the past, for products priced under $200, memory (DRAM + NAND) costs accounted for less than 20% of the total cost. Now, they exceed 40% or even higher than 50%. After factoring in pricing structures, if we apply the previous cost ratios, the very rationale for the product’s existence is lost. Moreover, this impact is interconnected – rising costs lead to higher recommended retail prices (RRPs), which in turn further influence consumers’ purchasing decisions. This pressure affects overall sales demand, and channels also see their revenue and profits impacted by selling fewer products, a situation that could lead to irreversible consequences. 

Years ago, this author was helping a company to sell solar solutions for African consumers – they had to pay mobile money installments for a product priced less than $50. So, if TECNO is telling such consumers that they have to pay $20 more for a product previously priced at $100, you will know how they will make the decision. 

Logical Fallacy of Subsidy Policy 

China’s nationwide trade-in subsidy policy is now in its second year, offering a 15% purchase subsidy (up to CNY 500) for smartphones priced under CNY 6,000, the same offer as 2025 but on a bit smaller scale. The original intent of this subsidy was to stimulate consumer electronics activity, encourage more people to upgrade to new devices, and get the consumption flywheel spinning. 

However, the situation has changed this year. Cost inflation has already pushed the pricing of high-end and flagship models, and even mid-to-high-end models, that Chinese OEMs had planned to expand, above the CNY 6,000 threshold. This means that once a brand’s product pricing crosses this line, it becomes ineligible for the subsidy. Besides, while lower-end and mid-range products appeared to offer better value for money last year due to the national subsidy, this year’s rising costs have effectively offset the benefits of the subsidy. OEMs are offering fewer product models and seeing lower sales volumes due to rising memory prices, leaving them with tighter cash flow and an inability to increase marketing investments to drive sales. Of course, the subsidy situation is a given, but if you look at the overall flow of funds, it doesn’t appear that the subsidies have stimulated more sales; rather, it seems that through this chain, the subsidy funds have flowed to memory suppliers, and what’s more, the impact of memory price hikes is even greater for products priced under CNY 6,000. 

Incompatibility of Budget Phones and AI Narrative 

We initially had high expectations for the GenAI smartphone market and defined it as follows: “A GenAI smartphone leverages large-scale, pre-trained generative AI models to create original content or perform context-aware tasks.” Even when discussing smartphones supporting 7 billion parameters, these devices mostly fall into the flagship or at least sub-flagship segments. In other words, the “AI narrative” does not encompass budget smartphones, despite claims that mid-range products can adopt a hybrid AI model – a claim that borders on sleight of hand. 

Industry insiders know that memory products follow a cyclical pattern. In this sector, three leading players from South Korea and the US control the majority of the global market, and they certainly need to align more closely with the AI narrative. Then there are the Chinese players, who are actually working hard; when it comes to budget smartphones, Chinese memory players are already the main force. However, in the early days, this led to cyclical problems: manufacturers invested heavily in expanding production lines and capacity. But once everyone’s capacity came online, overcapacity triggered price wars, causing prices to plummet rapidly. This time, AI presents manufacturers with an opportunity to change their fortunes: “I am already running at full capacity. If you want me to increase production and supply, fine, but you must sign a stricter long-term contract with me, ensuring you won’t drive down prices once capacity comes online.” Smartphone brands, however, hadn’t considered this scenario: “If I sign this contract, isn’t that effectively a death sentence for budget phones?” 

Even the industry’s major players – the carriers – have realized this. They originally intended to launch CPE devices, not to profit from them but to boost ARPU. Yet now, the proposals they are hearing from ODMs involve sourcing second-hand memory to meet supply needs, and at higher prices. 

The cheap smartphone market has always been driven by essential demand. In fact, the Chinese market has been relatively unaffected because it is currently undergoing a shift toward higher-end products. However, Chinese OEMs, which traditionally served global markets and met the needs of emerging markets, have faced a massive disruption to their supply chains due to the rise of AI.