The PC market’s latest decline raises a difficult question for manufacturers: how much cost inflation can they pass to buyers before those buyers postpone replacing a computer? Gartner’s preliminary third-quarter 2026 estimates, released October 12, put worldwide PC shipments at 55.4 million units, down 22.4% from a year earlier. Gartner attributes the contraction to higher memory and storage costs, increased device prices and a reduced supply of cheaper configurations. Its explanation points to affordability and product availability as central pressures on shipment volume.
All six leading vendors recorded declines. Lenovo shipped roughly 15 million units and retained a 27.1% share, with volumes down 22.4%. HP’s shipments fell 30.9%, Dell’s 25.4%, ASUS’s 13% and Acer’s 11.2%. Apple recorded the smallest decline among the six, at 7.1%, and increased its share to 10.5%. Gartner’s figures cover desktop and laptop PCs running Windows, macOS and Chrome OS. They measure shipments into distribution channels, are preliminary and may change when final estimates become available.
A higher average selling price can make a vendor’s financial picture look different from its unit trend, but it cannot be treated as proof that the business is healthier. Revenue depends on both price and quantity, while margins depend on how much of the price increase is absorbed by suppliers and other costs. If customers move toward expensive configurations because cheaper options are scarce, a richer mix may coexist with weaker access to the market. Investors and suppliers need to examine those effects separately rather than interpreting higher prices as stronger demand.
The replacement decision is particularly important. Buyers generally have some discretion over when to replace a functioning PC, even if they would prefer a newer machine. A price increase can move that purchase into a later quarter. Businesses may also direct a limited budget toward specific workloads instead of renewing an entire fleet. Those behaviors can produce a sharp shipment decline without implying that computers have become less necessary. The relevant question is whether purchases are deferred, redirected or lost for a longer period.
Apple’s relative performance illustrates another distinction: gaining share in a contracting market does not require selling more units. A vendor can lose volume more slowly than its competitors and emerge with a larger percentage of the total. That is meaningful competitive performance, but it should be described precisely. Similarly, Lenovo’s unchanged share indicates that its reported decline broadly matched the overall market. The headline ranking alone tells less than the relationship between each company’s shipment change and the industry’s contraction.
Affordability also changes what a household expects its home computer to be. A compact desktop such as the Mac mini, or a small Windows or Linux PC, can provide a permanent workstation while reusing a monitor, keyboard and mouse the owner already has. That separates the purchase of computing power from the purchase of an entire portable setup. It can make particular sense for someone who uses a smartphone on the move and wants a larger screen at home for writing, spreadsheets or editing photographs. The Mac mini remains a PC within Gartner’s macOS coverage: choosing a smaller enclosure does not make that purchase disappear from the shipment count. The significance is the buying decision and the equipment reused, rather than a new category outside the statistics.
At the lighter end, Raspberry Pi computers illustrate how little hardware some tasks require. A Linux desktop used for basic documents, modest browsing, coding experiments or a small home server need not resemble a full-sized consumer tower. There are compromises, though. The board alone is not the complete cost: storage, power, a case, cooling and peripherals may be needed, and performance or application compatibility can limit its usefulness. A Raspberry Pi is therefore one possible response to a narrow workload, rather than an interchangeable substitute for a Mac mini or a demanding photo-editing workstation. It also sits outside the Windows, macOS and Chrome OS definition used in the cited Gartner estimates.
The smartphone puts a different kind of pressure on replacement demand. Messaging, banking, shopping, navigation, casual photography, video calls and entertainment can all be handled on the device people already carry. For some households, the computer is becoming an occasional tool for tasks that benefit from a keyboard, more screen space or specialized software. When it is used less frequently, keeping an older machine for another year can feel entirely reasonable. Smartphones do not do everything equally well, especially sustained document work, complex spreadsheets and professional production. But they can do enough of the everyday work to weaken the urgency of buying another PC.
AI tools may also make Linux a more practical choice for users who previously found it too demanding. An assistant can explain an error message, help interpret a configuration file or suggest how to install software, reducing the effort involved in learning an unfamiliar system. Terminal assistants such as GitHub Copilot CLI support Linux, while browser-based AI services do not generally require Windows or macOS. Together, those options can help a user keep suitable existing hardware useful without buying a new machine simply to remain within a familiar software environment. This is a plausible adoption mechanism, rather than evidence that AI has already produced a measurable surge in Linux desktop use. Incorrect commands, unsupported applications and hardware compatibility still require care and judgment.
Linux introduces an additional measurement issue. Gartner’s stated operating-system coverage excludes Linux-only PCs, but installing Linux on an existing computer does not reverse the original shipment or create a new one. Its potential effect is on the next purchase: extending a machine’s useful life could delay replacement. The opposite effect is possible too, as AI development, local model experimentation and other demanding workloads create reasons to buy faster hardware. The market can therefore experience enthusiastic adoption of AI services alongside weaker mainstream PC shipments. More software activity does not automatically translate into more new computers.
The wider analytical context is a home-computing market with more ways to avoid an expensive replacement: keep the existing PC, use a compact desktop with existing peripherals, move light tasks to a small Linux machine or rely more heavily on a smartphone. AI could strengthen some of those alternatives by reducing software friction, even as it creates demand for powerful systems elsewhere. Gartner’s release does not quantify those substitution effects; its explanation for the reported quarter centers on component costs, prices and product availability. The longer-term question is how those immediate pressures interact with changing habits. Vendors need to sell a compelling reason to replace the machine, not simply demonstrate that the next model is faster.
The shipment estimates and vendor figures cited in this article are from Gartner’s preliminary worldwide PC shipment release dated October 12, 2026.
