Smartphones are undergoing a shakeout.

Gartner and IDC have both predicted that smartphones are entering a contraction era. Gartner forecasts an 8.4% decline in global smartphone shipments by 2026, while IDC predicts a 12.9% drop.

Amid this downturn, mid-range smartphones, once the backbone of the market, are becoming the biggest victims.

In recent years, the market share of mid-range smartphones has been eroded: low-end models with steadily improving quality are eating into their base, while high-end models with price reductions are delivering a dimension-reducing blow. Caught between these two forces, the share of mid-range smartphones has slipped from dominating half the market to less than 30%.

As if things weren't bad enough, in 2026, memory price hikes hit like a storm, prompting manufacturers to slash orders, with most cuts targeting mid-range and low-end models.

The underlying logic is not hard to grasp: mid-range smartphones offer poor profit flexibility.

Under the industry's 'reverse pricing' logic, high-end models can command premium prices through 'spec stacking,' providing enough profit cushion to absorb cost fluctuations.

In contrast, mid-range models have rigid cost structures, with a high proportion of 'non-negotiable' expenses. This results in memory chips accounting for more than double the cost in mid-range models compared to high-end ones. During memory price cycles, reducing mid-range models thus becomes a rational choice.

From a broader perspective, the decline of mid-range smartphones is merely a microcosm of the K-shaped era.

/ 01 / The Mid-Range Market Under Siege

Earlier this month, Meizu acknowledged that its domestic smartphone hardware R&D projects would be suspended, marking the end of an era for the brand.

Meizu's exit reflects the broader cooling of the mid-range smartphone market.

Mid-range smartphones, typically priced between 2,000 and 4,000 yuan, offer 85%-90% of the core experience of flagship models while cutting corners in non-core, high-cost areas. Their primary target audience consists of users with moderate spending power who prioritize practicality.

This positioning once allowed them to capture the largest common denominator of users, accounting for half of the domestic market at their peak.

How glorious they once were, how desolate they are now.

In recent years, the market share of mid-range smartphones has been on a downward trend. Goldman Sachs data shows a significant decline from 35% in 2021 to an expected 23% by 2027.

Even the 2025 national subsidies failed to reverse the decline of mid-range models.

The mid-range price segment was the most heavily covered by the subsidies, yet sales did not surge accordingly. According to Canalys and Counterpoint data, sales grew by about 4%-7% year-on-year, but growth was concentrated at the high and low ends of the market.

The growth of low-end models is not hard to understand. Emerging consumer giants like Mixue Ice Cream & Tea and MinMin have demonstrated the trend toward cost-effective consumption, ensuring growth for budget-friendly smartphones.

As for high-end models, their growth mirrors that of luxury goods during economic downturns. High-value users are unlikely to downgrade due to macroeconomic conditions.

Moreover, price reductions by high-end models are also siphoning market share from mid-range devices.

For instance, the iPhone 16 Pro was included in the national subsidies for the first time during last year's 618 shopping festival, driving Apple's monthly transaction volume up by nearly 200% month-on-month. During the Double 11 festival, the previous-generation iPhone 16 series saw price drops exceeding 20%, sparking a trend where buying mid-range models seemed less appealing than purchasing older Apple models.

Entering 2026, mid-range smartphones face new challenges on the supply side.

/ 02 / Memory Prices Continue to Pressure

"Memory prices have risen too much. We hope everyone understands our sincerity."

While Lei Jun tried to dampen backlash over the price increase of the Redmi K90 standard edition on social media, he also highlighted the new challenge for mid-range models: the dilemma of rising costs and squeezed demand due to memory price hikes.

This stems from the era's 'dark dividend' brought by AI. Today, the bottleneck in AI inference has shifted from 'insufficient computing power' to 'insufficient storage capacity.'

In short, if storage system performance lags, data cannot be delivered to GPUs and CPUs in a timely manner, leading to idle computing power as GPUs wait for data, ultimately causing AI applications to lag and efficiency to decline.

This phenomenon has triggered an explosion in AI storage, particularly HBM, which perfectly meets AI's high-bandwidth and low-latency demands. The global HBM market grew by 300% year-on-year in 2025, boosting both demand and profits. While Samsung's DRAM for traditional consumer electronics has an operating profit margin of 40%, HBM's margin reaches 60%.

Capital naturally flows to the most profitable areas. With limited capacity and surging AI storage demand, storage manufacturers like SK Hynix and Samsung have redirected over 40% of their DRAM (used in consumer electronics) capacity to HBM production.

This has squeezed consumer electronics storage capacity, leading to price hikes. After storage prices doubled in 2025, some institutions predict further increases of 40%-50% in Q1 2026 and another 20% in Q2.

As the saying goes, 'the consumer bears the cost.' Rising memory prices have triggered a wave of smartphone price hikes. However, higher prices also lead to shrinking demand, prompting many smartphone manufacturers to reduce their annual orders for complete devices. According to Jiemian News, the reduced orders primarily target mid-range, low-end, and overseas products.

The decision to cut mid-range and low-end shipments is understandable, given their high rigid costs and poor profit flexibility.

Specifically, consumer electronics like smartphones mostly adopt 'reverse pricing': setting the selling price and profit margin first, then allocating costs. High-end models rely on 'spec stacking' for premium pricing, offering ample profit margins and fault tolerance. Low-end models, on the other hand, require extreme cost-cutting, with a high proportion of 'non-negotiable' expenses in their cost structures.

Memory is a rigid cost for smartphones, accounting for an even higher proportion of costs in mid-range and low-end models. Memory chips directly account for 20%-34% of the cost in these devices, more than double the proportion in high-end models.

Thus, when faced with rising component costs like memory, high-end models can rely on profit cushions, while low-end models, lacking room for maneuver, see their profits directly eroded. Reducing mid-range smartphone production is therefore a rational choice for manufacturers.

The decline in mid-range smartphone market share is a microcosm of K-shaped consumption.

/ 03 / The Arrival of the K-Shaped Era

In 2020, U.S. financial analyst Peter Atwater proposed the concept of the 'K-shaped economy.' The shape of the letter K accurately describes a new phenomenon: after an economic shock, recovery does not follow a V-shaped trajectory where all sectors rebound together. Instead, the path forks—one branch rises, the other falls.

Beyond academic explanations, the public may resonate more with the real-world consumption divergence: affluent shoppers line up at luxury stores, while many supermarkets struggle to stay afloat. Pop Mart sees overwhelming demand, while baijiu (Chinese white liquor) stores sit empty.

The backdrop of K-shaped divergence is that we are at the tail end of an economic cycle and the turbulent beginning of a technological cycle. The overlap of these two cycles has not yet brought widespread prosperity but has instead ushered in an era dominated by 'stock economy.'

In a stock economy, high-net-worth individuals enjoy dividends from resource and technological monopolies, while the middle class and the general public lose opportunities for upward mobility, making 'divergence' inevitable.

This divergence will first manifest in the disappearance of middle-tier consumption, as seen in Japan's example.

During the Heisei era, the average salary of office workers declined, widening the wealth gap and accelerating the polarization of the middle class into two extremes, giving rise to an 'M-shaped' society and triggering a trend of consumption stratification. Players catering to both ends of the consumption spectrum achieved counter-trend growth during this period, with Kao Corporation thriving in the high-end market and Uniqlo in the cost-effective segment.

The K-shaped trend is not limited to class-based consumption divergence but also reflects shifts in individual consumption behavior.

When the 'elevator of the times' is ascending, users' 'investment/effort behaviors' yield relatively high returns. At such times, users prioritize practicality and value, making practical consumer goods core assets.

However, as the elevator slows, investment returns diminish, affecting everyone's mindset: when effort no longer guarantees reward, action begins to wane. In this context, 'emotional fulfillment' beyond practical functions provides a release valve for consumption, explaining the rapid rise of Pop Mart.

Confining the K-shaped era to the consumer sector would be myopic.

The emergence of new technologies has also driven industrial divergence. PwC's Global AI Jobs Barometer report, covering six continents and analyzing nearly one billion job postings and thousands of corporate earnings reports, also found traces of K-shaped divergence: industries with the highest AI penetration (financial services, software) saw productivity growth nearly quadruple since ChatGPT's release in 2022. In contrast, industries least affected by AI (mining, hospitality) experienced a slight decline in productivity growth over the same period. Industries with high AI exposure saw per-capita revenue growth three times that of other sectors.

The K-shaped trend has permeated every level of consumption, employment, and assets, albeit to varying degrees.