Smartphone makers' shipments fell 20%, and AI cannot be blamed for this

[[**In the third quarter of 2026, global smartphone shipments fell about 20% year-on-year, but this decline mainly stemmed from manufacturers' contraction of shipments to channels, rather than a simultaneous decline in consumer purchases. Channel inventory, memory price increases squeezing the profits of low-end models, and lengthening replacement cycles are the deeper reasons.**

**Key points**

**1. A 20% shipment decline cannot be directly equated with end-user sales**

Manufacturer shipment statistics count the number sent to channels, which is not equal to actual consumer purchases. In the first half of this year, manufacturers stocked up in advance due to expectations of memory price increases; after channel inventory accumulated, they began to actively reduce shipments, amplifying the statistical decline. Counterpoint data shows that during the same period, China's end-user sales decline was far lower than the channel shipment decline.

**2. Lengthening replacement cycles are a longer-term problem for the phone industry**

Phones are becoming more durable, and consumers' willingness to replace them continues to decline. Global shipments fell from about 1.505 billion units in 2018 to about 1.255 billion units in 2025, and the global average replacement cycle has lengthened to about 43 months. The old phones in users' hands run smoothly, and "the new model is more advanced" alone is already hard to persuade consumers to pay.

**3. Memory price increases hit low-end models far harder than high-end ones**

Profits on thousand-yuan phones are extremely thin; an increase of a few dozen yuan in memory cost can eat up most of the profit, so manufacturers actively reduce the supply of low-priced models. In Southeast Asia in the second quarter of 2026, among shipments of phones under $100, Xiaomi fell 69%, vivo fell 88%, and OPPO fell 96%. High-end brands such as Samsung and Apple saw relatively limited declines.

**4. Rising average prices do not equal successful premiumization; they may simply reflect a shrinking low-end business**

Xiaomi's phone shipments fell from 42.4 million units to 31.2 million units in the second quarter of 2026, and the average selling price rose from 1,073.2 yuan to 1,351.0 yuan, but phone revenue still fell 7.5% year-on-year. The increase in average price was mainly the result of reduced supply of low-end products, not enhanced competitiveness of high-end models.

**5. Cost relief can bring a short-term rebound, but long-term growth still needs a reason to upgrade**

After the supply and demand of memory chips recover, low-priced models may be relisted, and some consumers who held off because of prices will return. But what truly determines long-term sales is still whether phone manufacturers can use new technologies such as AI to create experiences that make users feel it is "worth upgrading," rather than simply relying on cost improvement. 2026-10-16 10:32 The troubles of phone manufacturers cannot be blamed on AI. Quick overview by Fangleng Tude La

This article is from the WeChat public account: Fangleng Tude La Shuzhai, author: Fangleng Tude La

In the third quarter of 2026, the global smartphone market delivered an ugly report card. A report published by Counterpoint shows that global smartphone shipments fell about 20% year-on-year, with some mainstream Android brands seeing even more obvious declines.

The market quickly found an explanatory path: AI data centers are snapping up memory chips, chip suppliers are allocating more capacity to more profitable servers, and phone manufacturers' procurement costs have risen accordingly. Phones become more expensive, promotions decrease, consumers delay purchases, and this ultimately shows up in sales. Phone manufacturers have also begun to talk frequently about memory cost pressure.

On one side are constantly rising memory prices, and on the other are suddenly falling phone shipments; the two things look like a perfect fit. But what actually happened in the phone industry? To answer this question, it is probably necessary to first understand how the surprising 20% in the news actually came about.

Phone sales plunged 20%—who exactly stopped buying?

Suppose a phone manufacturer shipped 1 million phones to dealer warehouses in the third quarter last year and only shipped 800,000 this year. According to industry statistics, shipments fell 20%. But whether the phones shipped last year were sold out, and how much inventory is still sitting in warehouses this year—this number will not tell us.

The 20% in the news belongs to the measure of manufacturers' supply to sales channels. How many phones consumers actually buy from stores and e-commerce platforms requires another set of statistics. In normal years, the two numbers generally move in the same direction; when there is concentrated stocking, price-cut inventory clearance, or changes in the timing of new phone launches, the gap between them may widen.

This year happens to have a special background: memory is getting more and more expensive. In the first half of the year, some manufacturers, worried that buying later would be more expensive, purchased components in advance and pushed goods into channels. For channel merchants, having goods on hand is not necessarily a bad thing, especially when the next batch of phones is expected to raise prices. Thus, even if consumers' enthusiasm for buying had already cooled, manufacturers' shipment books might not immediately look ugly.

Counterpoint's Q2 2026 China market tracker provides a ready example: channel shipments fell only slightly year-on-year, while actual consumer purchases had already shown a double-digit decline. Different versions for the same period had a difference of about 1% to 2% in the shipment decline, but pointed in the same direction: warehouses can receive phones noticeably faster than consumers buy them.

When the phones purchased earlier need to be digested, the situation reverses. Dealers are unwilling to take too much more goods, and manufacturers also begin to arrange production cautiously. At the same time, component costs continue to rise, and even if some low-priced models can be sold, their profits may be so thin that there is no need to continue distributing them. Reducing shipments thus became a business choice actively made by companies.

Of course, it cannot therefore be said that consumers were completely unaffected by price increases. The phone end-market did indeed see a decline in purchases this year, and reduced promotions did indeed make people who had originally planned to upgrade continue to wait and see. For a phone that rose from 1,999 yuan to 2,299 yuan, for budget-limited users, the reason is already sufficient.

Counterpoint statistics show that during the four weeks of the 2026 618 promotion period, China's smartphone end-user sales fell 13% year-on-year. This time it is real consumer purchases; rising costs limited manufacturers' promotional efforts. It proves that the impact of price increases is real, and also reminds us not to treat end-user sales and the third-quarter global shipment decline as the same indicator.

The problem lies in the next inference: consumers buying fewer phones because of price increases, and the entire industry shipping 20% fewer phones to channels, cannot be directly equated. This year's cost crisis has simultaneously changed consumers' choices and manufacturers' shipping plans, with both directions acting on the market together. If these accounts are not separated first, that 20% can easily be attributed entirely to consumers.

Phones are getting more durable, and manufacturers are feeling worse and worse

More than a decade ago, changing phones often meant changing a lifestyle. People who had used feature phones picked up a smartphone for the first time and could install all kinds of apps; moving from 3G to 4G suddenly made video, maps, and mobile payments easy to use; each time screens, cameras, and fingerprint recognition crossed a threshold, users could feel the change.

New phones back then, even if a bit more expensive, made it relatively easy for manufacturers to tell consumers why they were worth buying. Old phone not broken yet? No problem, the new phone has features you couldn't use before. The changes shown at the launch event could be experienced directly in daily life the next day.

Now, the most awkward competitor for phone manufacturers is their own selves from three or four years ago. Take an old flagship: WeChat, Douyin, Bilibili, navigation, food delivery, and QR code payments still work. The camera may not be as strong, and game frame rates may not be the highest, but the functions ordinary people truly need every day were long ago fulfilled almost completely.

Of course, phones have not stopped improving. Processors are faster, imaging algorithms are better, screens are brighter, and foldable screens are starting to mature. But how much technical parameters have improved and how many thousands more consumers are willing to spend are two different accounts. The benchmark score improvements emphasized at launch events, when they reach users' hands, may just mean opening an app used to take one second and now takes a little less waiting. For most people, it is really hard to walk into a phone store specifically to save that little bit of time.

This change directly affects replacement frequency. If there are 100 million users, each changing phones once every three years on average, that creates replacement demand of more than 30 million units per year; if everyone uses their phone one more year, with the number of users unchanged, the corresponding replacement demand each year drops to only 25 million units. Manufacturers did nothing wrong; simply because old phones live one more year, the new orders they can compete for shrink by a large margin.

Counterpoint's annual data presents a fairly clear curve: global shipments were about 1.505 billion units in 2018 and about 1.255 billion units in 2025. There were rebounds in between, but the endpoint is still lower than the starting point. Another replacement cycle study shows that the global average replacement cycle reached 43.4 months in 2023 and about 43 months in 2024.

Figure 1 | Long-term changes in global phone shipments. Mature markets still rebound, but the basis for growth is already different.

Some may ask, if the market is so mature, why did global phone shipments rebound in the past two years? The reason is not mysterious. Phones purchased during the pandemic gradually aged, some regions saw concentrated replacements, and brands such as Apple also attracted a batch of users with new products. Mature markets will still grow, but increasingly need a specific reason to buy; it is hard to expect users to proactively spend money every fixed two years.

For consumers, this is actually a good thing. Electronic products that cost three or four thousand yuan in the past can still be good after four or five years, so the average annual cost of use naturally declines. Phone manufacturers spent more than a decade making products more and more durable, and finally discovered that consumers really are reluctant to replace them. Product managers can be proud of this, but the sales department probably cannot smile.

Phone manufacturers must find a reason worth paying for the next phone. To sell the next phone, it is already hard to rely only on telling users that "the new model is more advanced." They also have to make users believe that the advanced part happens to be worth their money.

Old phones have protected wallets, and new memory price increases have blocked the retreat of low-priced new phones. As cost pressure reaches phone manufacturers, they begin to recalculate: which phones are still worth continuing to sell?

Once memory prices rise, cheap phones are the first to be dissuaded

The thousand-yuan phone market of the past was much like a race with no rest days. Today someone brings a large battery below a thousand yuan, and tomorrow a competitor adds faster charging and more memory. Consumers are certainly happy to see this; as long as they wait a few more months, the same money always seems to buy a phone with better specs.

But the profit a thousand-yuan phone leaves manufacturers is far less ample than the selling price suggests. Screen, processor, memory, camera, battery, assembly, transportation, channels, and after-sales all take a share. Manufacturers also bear R&D and marketing expenses, and how much profit can be kept per unit requires careful calculation.

Let's do a simple hypothetical account: a phone sells for 1,000 yuan, and after deducting related costs, suppose it could originally earn only 50 yuan. If memory costs increase by 40 yuan, profit shrinks from 50 yuan to 10 yuan. Costs increase by only 4% of the selling price, yet eat up 80% of the original profit. If costs rise a little more, the original business may no longer be worth continuing.

Counterpoint's estimate for a typical low-end phone with 6GB of memory and 128GB of storage shows that in the first quarter of 2026, memory and flash storage costs account for about 43% of the total component cost, and the model's total component cost increases by about 25% quarter over quarter. This is not the average for all thousand-yuan phones, but it is enough to explain why low-priced products are especially hard to endure.

When low-priced models barely make money, reducing supply also becomes a business choice. At this point manufacturers have only a few options. Keep the price unchanged and swallow the cost; pass the price increase to consumers and risk declining sales; cut specifications, but easily lose to competitors in comparisons. There is another way: simply reduce models with excessively low profits and leave production capacity, channel resources, and marketing budgets for more expensive products.

This change has already appeared in overseas markets. Some Chinese phone brands have reduced products in extremely low price segments and pushed up the starting prices of entry-level models. Of those vanished sales, part is consumers unwilling to pay for price increases, and another part is manufacturers proactively reducing supply of such low-profit products. The two affect each other: with fewer choices, buyers are more likely to keep using old phones; when users buy less, manufacturers are even less willing to maintain the original supply scale.

According to Omdia data, in Q2 2026 Southeast Asia shipments of phones under $100, Xiaomi fell 69%, vivo fell 88%, and OPPO fell 96%. Higher price segments did not fully absorb the vanished low-price demand, meaning some purchases left the new-phone market entirely.

Figure 2 | Southeast Asia low-price phone supply contracts. The above are year-on-year declines for brands in the specified price segment.

In the same round of cost shock, Apple, Samsung and some high-end models were affected differently. The reason is easy to understand. For a thousand-yuan phone, an added cost of a few dozen yuan may swallow most of the profit; for a product priced five or six thousand yuan, companies usually have more room to adjust. Brands have bargaining power and inventory capability, and consumers may also value ecosystem, usage habits, imaging performance and after-sales service more.

Counterpoint statistics show that in Q3 2026 Samsung's global shipments fell only about 3% year-on-year, and Apple's market share reached 22%, although Apple's own shipments were also declining; Xiaomi's shipments fell more than 30%, and OPPO and vivo also more than 20%. Brands differ in supply, models and launch timing, so this set of data alone cannot establish that high-end consumers are completely unafraid of price increases.

But the relatively better performance of high-end brands does not imply that wealthy consumers do not care about price. If one manufacturer keeps its original price and another raises a comparable product by several hundred yuan, users who were originally hesitant may instead lean toward the former. Consumers are always comparing how much value they get back for the money they spend.

A rising average selling price does not mean every phone is raising prices. Suppose a manufacturer sold ten 1000-yuan phones and ten 3000-yuan phones last year, and this year sells only five of the cheap ones while still selling ten of the expensive ones. Even if neither phone rose by a single cent, the calculated average selling price would still increase. At the same time, total sales fell by a quarter. This number sounds like successful premiumization, but in reality it may also be a shrinking low-end business.

Xiaomi's Q2 2026 announcement provides a real operating comparison: phone shipments fell from 42.4 million to 31.2 million units, while the average selling price rose from 1073.2 yuan to 1351.0 yuan. At the same time, phone revenue still fell 7.5% year-on-year. A higher average price may not offset the revenue loss caused by selling fewer phones.

Table 1 | Xiaomi Q2 2026 phone business: shipments and average price move in opposite directions.

What memory price increases change is not only how much consumers have to pay, but also which price segments manufacturers are willing to serve. For people with limited budgets, this change is especially painful: phones that were once affordable are now more expensive, and cheap models that once filled the shelves are gradually becoming fewer. So old phones can be used for another year.

When memory gets cheap again, who will be responsible for making people want to change phones?

Suppose memory chip supply gradually recovers in the future and manufacturers can again purchase memory at reasonable prices. The first to be happy should be the product managers making entry-level phones. Cost calculations that previously did not add up suddenly add up again; shelved low-price models can be relisted, and distributors are willing to restock.

Consumers will benefit too. People preparing to change phones find prices loosening and may stop waiting; models that were previously over budget re-enter their choice set after price cuts. If channels also begin restocking, manufacturers' shipments may even recover faster than consumers' actual purchases. After a period of time, industry reports may look much better than this year.

IDC once forecast that as storage supply and demand gradually recover, global smartphone shipments in 2028 could rebound 5.5%. This forecast still carries great uncertainty, but it reminds us that easing cost pressure may indeed bring a substantial recovery, and such a rebound cannot be dismissed in advance.

How much short-term sales can recover depends on why consumers originally did not buy. One person's phone already crashes frequently and they would pay if only the price dropped a few hundred yuan; cost recovery is very useful to them. Another person's phone runs smoothly and the battery still lasts a day, and they have never seriously considered changing phones; even if the same new model is 500 yuan cheaper, it may not change their mind.

Price cuts can call back those waiting for prices, but it is hard to wake someone who has no plan to change phones at all. Looking at these two types of demand together, the short-term recovery and long-term growth of the phone industry have a clear boundary. The former can rely on cost improvement, promotions and restocking; the latter still needs products to recreate the necessity of purchase.

What phone manufacturers are now betting on as the new reason happens to still be AI. From voice assistants to phone agents that automatically complete operations for users, manufacturers hope the next phone change will once again bring a perceptible leap in experience. This is indeed possible. In the past, mobile payments and the mobile internet changed how phones are used, and new interaction technologies may also have a chance to create demand.

However, whether users are willing to change phones also depends on whether this capability can truly enter daily life. If AI merely shortens a few steps that previously had to be done manually, and old phones can get a similar experience through app updates or cloud services, users may not be willing to buy new devices for it. If in the future certain features must rely on new hardware and can continuously save time and solve problems that previously could not be solved, the reason to buy will become solid.

This is also the most ironic scene in the phone industry this year: AI has taken away memory chips, making phones more expensive to manufacture; manufacturers turn around and expect AI to find the next round of growth momentum for phones. AI adds trouble for the production department on one side while also having to deliver performance for the sales department on the other; the work schedule is really a bit tight.

By then, the normalization of the supply chain may bring back some sales, but the real test still lies before phone manufacturers. Cheaper memory can make new phones cheap again, but it cannot suddenly make old phones hard to use.

Only cost increases always have a chance to ease. When phones become cheap again, launch events will still be lively, and new product specs will still look better year after year. As for whether consumers will pay, it probably still depends on whether the old phone in their hand agrees. Channel: 3C Digital

73 articles published on Huxiu. This content is authorized by the author for publication; the views represent only the author and do not represent Huxiu's position. If you have objections or complaints about this manuscript, please contact tougao@huxiu.com. Everyone who is changing and wants to change the world is on the Huxiu APP[ ]]]