Hewlett-Packard has issued a stark warning to investors and industry observers: the cost of memory has ballooned to such an extent that it now represents roughly 35% of the total bill of materials for its personal computers. The disclosure, made during a recent earnings call, underscores a growing crisis across the PC industry as DRAM prices continue their relentless upward march, squeezing margins and threatening to reshape how manufacturers design and price their machines.
The revelation came from HP CEO Enrique Lores, who told analysts that memory costs have become the single largest component expense in the company’s PC lineup. According to reporting by TechRadar, Lores characterized the situation as a significant headwind, noting that the proportion of cost attributable to RAM has climbed sharply compared to historical norms. In previous years, memory typically accounted for a far smaller share of overall PC component costs, often ranging between 10% and 20% depending on the configuration.
A Perfect Storm of Demand and Supply Constraints
The surge in DRAM pricing is the product of several converging forces. The explosive growth of artificial intelligence workloads has created enormous new demand for high-bandwidth memory, particularly HBM (High Bandwidth Memory) chips used in data center GPUs. While HBM and standard DDR5 DRAM are different products, they compete for the same manufacturing capacity at major memory producers like Samsung, SK Hynix, and Micron. As these chipmakers have shifted production lines toward the more lucrative AI-oriented memory products, the supply of conventional PC DRAM has tightened considerably.
At the same time, the transition from DDR4 to DDR5 memory — now standard in most new PCs — has introduced its own cost pressures. DDR5 modules are inherently more expensive to produce due to their more complex architecture, which includes on-die voltage regulators and higher-density chip designs. The combination of constrained supply and a more expensive baseline product has created what HP’s leadership describes as an unprecedented cost challenge.
AI PCs Are Making the Problem Worse
Adding fuel to the fire is the industry’s aggressive push toward so-called “AI PCs” — machines equipped with dedicated neural processing units (NPUs) and substantially more RAM to handle on-device AI inference tasks. Microsoft’s Copilot+ PC specification, for example, recommends a minimum of 16GB of RAM, with many manufacturers opting for 32GB or even 64GB configurations to ensure adequate performance for local AI models. These higher memory requirements directly amplify the cost impact of rising DRAM prices.
HP itself has been at the forefront of the AI PC movement, launching multiple product lines designed to run AI workloads locally rather than relying solely on cloud processing. But as Lores acknowledged, this strategic direction comes with a tangible financial burden. Every additional gigabyte of RAM in an AI-optimized laptop or desktop translates directly into higher component costs that must either be absorbed by the manufacturer or passed along to consumers.
The Ripple Effects Across the PC Industry
HP is far from alone in grappling with this challenge. Dell Technologies, Lenovo, and other major OEMs are facing identical pressures, though HP’s candid public acknowledgment of the scale of the problem has drawn particular attention. Industry analysts at firms like TrendForce have been tracking DRAM contract price increases throughout 2024 and into 2025, with some categories seeing quarter-over-quarter price hikes in the range of 15% to 20%.
The implications extend beyond just the sticker price of a new laptop. Corporate IT departments, which purchase PCs in bulk and operate on tightly managed budgets, may be forced to delay refresh cycles or opt for lower-memory configurations. Educational institutions, government agencies, and small businesses — all price-sensitive buyers — could find themselves priced out of the latest hardware at precisely the moment when AI capabilities are becoming a competitive necessity.
Memory Makers Are Prioritizing AI Over PCs
From the perspective of DRAM manufacturers, the current pricing environment is a welcome reversal after years of cyclical downturns that hammered profitability. SK Hynix, which dominates the HBM market, has reported record profits driven by insatiable demand from Nvidia and other AI chip companies. Samsung and Micron have similarly benefited, and all three major producers have signaled that they intend to continue prioritizing high-margin AI memory products over commodity PC DRAM.
This strategic calculus by the memory industry creates a structural problem for PC makers. Unlike a temporary supply shock caused by a natural disaster or factory outage, the current DRAM shortage is driven by a deliberate reallocation of manufacturing capacity toward more profitable product categories. Until memory producers invest in significant new fabrication capacity — a process that typically takes two to three years from groundbreaking to volume production — the supply-demand imbalance for PC-grade memory is unlikely to ease substantially.
HP’s Strategic Response and Margin Management
In response to the cost pressures, HP has indicated it will employ a multi-pronged strategy. The company plans to optimize its product configurations more carefully, potentially offering tiered memory options that allow price-conscious buyers to choose lower-RAM models while reserving premium configurations for users who need maximum AI performance. HP has also signaled that some cost increases will inevitably be passed through to end customers, though the company is attempting to manage this carefully to avoid dampening demand.
Lores also pointed to operational efficiencies and supply chain negotiations as partial offsets. HP, as one of the world’s largest PC manufacturers by volume, has significant purchasing power with memory suppliers. But even that leverage has limits when the underlying commodity is in short supply across the entire market. As TechRadar noted, the 35% figure represents a dramatic shift in the cost structure of a PC, effectively making memory the most expensive single component — surpassing even the processor in many configurations.
What This Means for PC Prices Going Forward
For consumers and enterprise buyers alike, the outlook suggests that PC prices are likely to remain elevated or even increase further in the near term. The convergence of AI-driven demand, the DDR5 transition, and constrained manufacturing capacity creates a pricing environment with few near-term relief valves. Some analysts have speculated that DRAM prices could begin to moderate in late 2025 or early 2026 as new fab capacity comes online, but the timeline remains uncertain.
There is also a question of whether the industry might explore alternative approaches to reduce memory dependency. Technologies like memory compression, more efficient caching algorithms, and hybrid storage architectures that use fast SSDs as virtual memory could help mitigate the need for ever-larger RAM pools. However, these are incremental solutions that cannot fully substitute for physical DRAM, particularly in AI workloads that require rapid access to large datasets held entirely in memory.
A Structural Shift in PC Economics
HP’s admission marks a significant moment for the PC industry. For decades, the cost structure of a personal computer was dominated by the processor, with memory and storage playing supporting roles. The inversion of that hierarchy — with RAM now commanding the largest share of component costs — represents a fundamental change in how PCs are designed, priced, and sold.
The broader lesson is that the AI boom, while generating enormous value in data centers and cloud services, is imposing real costs on adjacent industries. PC manufacturers are effectively subsidizing the AI revolution through higher memory prices, and those costs are ultimately borne by the hundreds of millions of consumers and businesses that buy new computers each year. How the industry adapts to this new reality — through pricing strategies, product design, and technology innovation — will be one of the defining challenges of the next several years.
For HP, the transparency of the disclosure is itself notable. By publicly quantifying the scale of the memory cost problem, the company is setting expectations with investors and signaling to suppliers that the current trajectory is unsustainable. Whether that message prompts any change in behavior from DRAM manufacturers remains to be seen, but the data point itself — 35% of PC costs consumed by memory alone — is likely to become a frequently cited benchmark in industry discussions for months to come.