AI’s Hidden Tax: Why Smartphones & PCs Are About To Get More Expensive

Welcome to Memorandum Deep Dives. In this series, we go beyond the headlines to examine the decisions shaping our digital future. 🗞️

This weekend, we’re diving into a consequence of the AI boom: why smartphones and PCs are getting more expensive, not because demand collapsed, but because AI infrastructure is outbidding consumer technology for the components that matter most.

For decades, consumer devices set the pace of the tech industry. Phones and laptops dictated chip roadmaps, factory expansions, and quarterly expectations. Today, that center of gravity has shifted decisively toward AI data centers, where chips run nonstop and memory is consumed at an industrial scale. As capital, capacity, and components flow toward AI workloads, consumer electronics are learning what it means to compete in an ecosystem they no longer lead. What looks like a pricing problem is really a structural one—and it’s only just beginning.

Why AI could raise the cost of smartphones and PCs

Not long ago, consumer tech set the rhythm of the tech industry. Smartphones and PCs dictated chip designs, factory capacity, and investment priorities. Product launches would dominate the news cycle, and at the end of each fiscal quarter, tech news outlets would report on device shipments and discuss growth plans.

Today, that story has changed. The center of gravity has moved to AI data centers, where value comes from processing enormous amounts of data and running models around the clock. Chips are no longer judged by how many end users they have, but by how much work they can handle. Although consumer technology still matters, it no longer leads the way.

A stark reminder of this silent shift is the fact that companies that once made headlines for reporting massive growth are now facing diminishing returns from manufacturing consumer tech devices.

Over the past week, numerous news articles have detailed how a simple component, the memory chip, has become the focal point of a broader transition that places AI data centers, rather than consumer electronics, at the heart of the tech industry.

Memory becomes the bottleneck.

On February 4, 2026, Reuters reported that shares of Qualcomm and Arm Holdings tumbled (down 10% and 8%, respectively) after their quarterly results.

On 6 January, Bloomberg reported that Samsung expects memory chip supply shortages to impact everyone. And the coverage does not stop there.

Earlier this month, Reuters reported that Apple said rising memory chip prices were beginning to pressure profitability and would have a greater impact in the current quarter.

The squeeze was driven by chipmakers shifting production toward high-margin AI memory, tightening supplies of conventional DRAM for phones and PCs.

Memory chip makers Samsung and SK Hynix further warned that the shortage would persist as AI demand took precedence, pushing up costs, squeezing margins, and weighing on device shipments.

This shift then directly affects the availability of memory chips for consumer electronics, an essential part of modern computer systems.

Most consumers today focus on the processor, to the extent that much of the marketing centers on faster chips, more cores, and improved performance benchmarks. Memory rarely makes headlines, yet it is the silent backbone of every modern computing system. Without memory, processors have nothing to work on and nowhere to store results.

RAM determines how many tasks a device can handle simultaneously and how quickly it responds, whereas storage chips govern how much data can be stored and how quickly it can be accessed. Every photo, app, AI model, and operating system depends on both. As semiconductor analysts frequently point out, a cutting-edge processor paired with insufficient memory is like a sports car stuck in traffic.

The power of the Big Three

Making matters worse is the dominance of three companies, which control approximately 94–95% of the global DRAM market (the memory used in PCs and phones) and the vast majority of the high-speed memory required for AI.

These are Samsung Electronics, SK Hynix, and Micron Technology.

These Big Three hold a near-total grip on the world’s supply because they have perfected the highly complex process of manufacturing high-speed memory at a massive scale. As of early 2026, they have pivoted their production lines to produce High-Bandwidth Memory (HBM), which is essential for AI chips sold by companies such as Nvidia.

Because HBM consumes significantly more space on a silicon wafer than standard memory, it effectively produces fewer chips overall. This has created a market vacuum for everyday electronics such as smartphones and laptops, allowing these companies to increase prices by 60%-70% as demand far outstrips supply.

Why memory shortages are hard to fix

Another problem plaguing the industry is the difficulty of addressing the memory chip shortage.

Unlike processor shortages, which can be fixed by shifting production between foundries or adjusting designs, memory shortages are far harder to work around.

Building or expanding memory fabrication plants takes years and tens of billions of dollars. According to the Financial Times, memory manufacturers have historically been burned by boom-and-bust cycles, leading them to expand cautiously even when demand surges.

Memory is also a lower-margin business than advanced processors. That makes suppliers more selective about where they allocate scarce capacity. When demand spikes suddenly, as it has with AI, there is little slack in the system. The result is not just higher prices but outright shortages that ripple through the supply chain.

This is why the current crunch cannot be waved away as a temporary disruption. It reflects long-term structural limits in how quickly the memory supply can respond to demand.

AI outbids consumer electronics.

Additionally, over the past couple of years, AI’s rapid growth has driven high demand, crowding out other areas.

AI workloads are fundamentally different from consumer workloads. Training large models requires massive amounts of high-bandwidth memory to move data quickly between processors. Running those models in production also consumes memory continuously, often around the clock. Reuters and Bloomberg have both reported that leading cloud providers are buying memory in volumes previously unseen in the industry.

Unlike smartphones, which are shipped once and then remain idle for much of the day, AI servers are designed for continuous operation. That makes them far more profitable customers for memory suppliers. When memory is scarce, suppliers naturally prioritize buyers who can pay more and commit to long-term contracts.

This is not a moral judgment. It is basic economics. AI data centers have become the most powerful magnets in the semiconductor ecosystem, pulling memory away from consumer devices even when those devices still have end-user demand.

The ripple effects on consumers

For smartphone makers, the choice is simple.

Add to this the razor-thin margins the smartphone makers often operate on compared to AI infrastructure, and the reason behind Apple’s recent earnings report becomes clear.

During its latest earnings report, Apple did not blame weak demand. The company said demand for its latest iPhone was ‘staggering.’ The real issue was cost. As memory makers shifted production toward AI-focused chips, Apple found itself paying more for the same DRAM used in iPhones and Macs.

Memory is not something you can easily replace or design around, so Apple was constrained. Either it absorbs the higher costs and watches its margins shrink, or it passes some of that increase on to customers.

That same squeeze is now rippling through the smartphone and PC industries.

PC makers are facing similar constraints, leading to delayed launches, limited configurations, or higher prices. Some manufacturers are quietly reducing specifications to conserve memory, offering lower-capacity models as defaults.

Consumers may not immediately notice why a device costs more or why certain configurations are unavailable. But over time, these constraints shape buying behavior and slow upgrade cycles. The result is a feedback loop where weaker consumer demand reinforces the industry’s shift toward AI-focused investment.

In this way, memory shortages are not just a supply problem. They subtly reshape what consumers expect and what companies are willing to build.

Even if manufacturers seek alternatives, the situation is unlikely to improve soon.

Reuters has reported that PC makers such as HP, Asus, and Dell are considering sourcing memory from Chinese suppliers to ease pressure.

This introduces new risks. Chinese memory chips may not always match the performance or reliability of established suppliers, and geopolitical tensions complicate long-term commitments. Yet scarcity forces difficult choices. When production lines are at stake, companies prioritize continuity over caution.

These decisions hint at a more fragmented future for technology supply chains, in which efficiency and resilience increasingly clash with political and security concerns.

Under AI’s shadow: The new normal for tech

Scroll through the technology section of any news outlet, and you will see the extent of AI’s impact on the technology sector. Today, reports indicate that fears of automation in the software sector wiped out nearly a trillion dollars in market value from software company stocks.

Memory shortages are exposing a change that has been building for years. As AI infrastructure absorbs more of the world’s chip capacity, consumer devices are learning to compete in an ecosystem they no longer dominate.