Memory Price Surge Is Forcing Smartphone Makers Out Again

The AI Tax on Your Next Phone

There is a new line item on every smartphone bill, and it has nothing to do with 5G modems or camera sensors. It is memory—DRAM and NAND—and its price is climbing because AI data centers are eating the supply that the smartphone industry used to take for granted.

Memory fabricators like Samsung Electronics, SK Hynix, and Micron are increasingly prioritizing high-margin AI server chips—HBM and enterprise-grade DRAM—over the commodity-grade components that ship by the billions inside phones. The margin spread between these two segments is stark: HBM can command premiums of 30 to 50 percent over standard consumer DRAM, and demand shows no sign of plateauing. The result is a slow but relentless squeeze on the low-end smartphone market that could force another wave of exits, echoing the earlier consolidation that drove Japanese makers out of the business.

This time, the danger zone is not limited to domestic players in any single country. ASUS has reportedly ceased new smartphone development and is effectively pulling out of the business in 2026 after last shipping ROG Phone models through 2025. Nothing Technology faced rumors of withdrawal from more than a dozen markets, including Japan; it denied the reports, but the story itself is telling—a brand that built its identity on design provocation and direct-to-consumer transparency is now sitting on supplier negotiations that increasingly feel like a negotiation with scarcity.

Who Wins, Who Loses

The winners are the giants. Motorola Mobility, now owned by Lenovo, and Xiaomi have scale advantages that translate into procurement power and the flexibility to shift production across product lines. They can absorb memory cost increases by squeezing margins on lower-tier devices or by raising prices incrementally and watching which segments stay viable. Samsung, its own memory manufacturer, operates in a category all its own: it is both buyer and seller, a position that insulates it from much of the turbulence.

The losers are everyone else. Niche brands like ASUS and Nothing, regardless of geography, lack the volume to negotiate favorable terms and the capital reserves to wait out a supply crunch. Their phone portfolios are thin, their cash flow narrow, and their exit costs are essentially just a press release away. What makes their situation especially precarious is that they cannot fall back on other revenue streams the way a diversified conglomerate can. When smartphone losses mount, there is no server division or cloud business to subsidize them.

The market structure is shifting toward oligopoly in ways that go beyond typical industry consolidation. This is not a cycle where the strong survive and the weak adapt. This is a cycle where the weak exit entirely—and with them goes the diversity of device design that keeps the market competitive on features, form factors, and pricing experiments.

The Consumer Ceiling

The core tension is simple: memory prices are rising, smartphone prices must follow, and consumers have a hard limit on what they will pay. When that limit is breached, purchase intent drops. Sales weaken. Makers with thin portfolios feel the revenue fall first because they cannot offset weak-performing lines with volume leaders.

The Asian low-end market—India, Southeast Asia, parts of Latin America—is especially exposed. Entry-level phones in these regions frequently sit at the ¥5,000 mark or below in domestic Japanese terms, and memory is a disproportionately large share of the bill of materials for those devices. A 10 percent swing in DRAM costs can wipe out an already thin margin or force a price jump that pushes a device past a psychological price barrier and kills demand outright. In markets where consumers trade in phones every 18 to 24 months and treat upgrades as discretionary rather than essential, even a modest price increase can collapse a model’s viability.

Chinese manufacturers facing the sharpest pressure include firms that compete primarily on price rather than brand loyalty. These are the companies that operate on single-digit margins and rely on volume to sustain profitability. When memory costs rise, they do not have the cushion that Samsung, Apple, or even Xiaomi possess. Some are attempting to reprice upward, but that strategy carries its own risk: in markets where brands like Realme, Redmi, and Poco have trained consumers to expect extraordinary value, raising prices can trigger a demand collapse that is harder to recover from than a margin squeeze.

Second-Order Effects

The consequences of this squeeze extend well beyond who shuts down and who doesn’t. For one, the exit of smaller players reduces competitive pressure on incumbents, which tends to soften innovation timelines and reduce the pace at which new features reach mainstream devices. When ASUS was actively competing in the gaming-phone niche, it pushed Motorola and others to respond with their own performance-focused hardware. Without that pressure, the segment risks stagnation.

There is also a supply-chain feedback loop to consider. Memory manufacturers that have pulled capacity away from consumer DRAM and NAND to serve AI demand are making capital-intensive bets on that trajectory. If smartphone exits accelerate rapidly, the resulting drop in consumer memory demand could create a future glut once AI capacity planning corrects—a scenario that would punish the very manufacturers trying to weather the current crunch. The industry is walking a narrow path between shortage today and oversupply tomorrow.

Carriers and retailers are feeling the effects too. In markets where subsidized phones drive plan adoption, the inability to source affordable hardware at previous price points is forcing renegotiation of carrier-brand agreements. Retailers are watching shelf space become a more contested resource as fewer brands compete for it, which further entrenches the advantage of those with distribution deals already in place.

And there is a geopolitical layer: as Chinese manufacturers face margin compression, some are likely to accelerate investment in in-house memory production or deepen partnerships with domestic foundries. China already treats semiconductor self-sufficiency as a strategic priority, and this pressure could accelerate the timeline. Whether that investment translates into meaningful supply for consumer electronics within a useful window remains an open question, but the direction of policy and capital is clear.

What Happens Next

If the memory squeeze persists through 2026 and beyond, expect another round of exits—not only from niche international brands but potentially from mid-tier Chinese manufacturers who have been operating on razor-thin margins. The pattern mirrors 2023, when domestic Japanese makers were forced out, but the geography of attrition is now broader and the competitive dynamics more globally distributed.

Two scenarios are plausible. In the first, memory supply catches up as manufacturers eventually expand capacity for consumer-oriented DRAM and NAND alongside their AI-focused lines. Prices stabilize, and the market resets around a smaller number of surviving brands. This outcome requires both sustained investment in fabrication capacity and a moderation in AI-driven demand growth—neither of which is guaranteed.

In the second, AI demand remains structurally dominant and memory stays tight for years. In that world, low-end smartphone prices settle at a higher floor permanently, and the gap between entry-level and mid-range devices narrows as manufacturers abandon the bottom tier entirely. Consumers who previously could buy a functional phone for under $150 may find that threshold moving to $200 or beyond, effectively pricing out a segment of the global market that has historically been served.

The second outcome is arguably more dangerous for the industry. Fewer participants mean fewer experiments, fewer new entrants, and less pressure on incumbents to innovate on price or features. The market stagnates from the bottom up. The segment that once functioned as a pipeline—bringing first-time smartphone users into ecosystems, eventually upgrading them to higher-margin devices—risks becoming a no-man’s-land where entry is too expensive for some and profit too thin for others.

The Bigger Picture

The memory shortage is not just a component-supply problem. It is a structural reallocation of semiconductor capacity driven by AI investment, and smartphones are losing that competition for silicon. The signal from this reallocation is clear: the industries that command the highest margins win the chips, and consumer electronics lose. Data centers, cloud providers, and AI platforms are willing to pay far more per gigabyte than anyone selling $200 phones. The market has spoken, and the allocation is efficient in the narrow sense—even if it leaves a significant portion of global consumers stranded.

For the global smartphone market, the implication is that the low-end segment—where growth in unit volume has been strongest—faces a cost crisis with no easy fix. Manufacturers that can navigate it will be the ones with scale, diversified product lines, and the willingness to let marginal businesses die. Everyone else will watch their market share erode not because of a better product from a competitor, but because the raw cost of making their product went up and they could not pass it on.

ASUS is the first major name to exit under these conditions. Nothing’s uncertainty is the next chapter. But the structural forces at work are not specific to any single brand or region. They reflect a fundamental realignment of semiconductor economics that is unlikely to reverse quickly. If history is any guide, ASUS will not be the last.