Goldman Sachs Cuts Global PC Shipment Forecast, Projecting 14% Decline in 2026
N.R. Finch
Goldman Sachs slashed its 2026 global PC shipment forecast from a 10% decline to 14%, citing rising memory and CPU costs plus a fading Windows 10 upgrade cycle; for investors, the PC hardware cycle trough is now later and deeper than previously expected.
How deep is the cut?
Goldman's revised numbers: 2026 global PC shipments at 255 million units, down 14% year-on-year; 2027 at 243 million, down 5%; 2028 at 244 million, flat.
This means → not a single year of growth across the three-year window. The prior forecasts of +3% for both 2027 and 2028 have been entirely withdrawn.
In plain terms = Goldman previously expected a bottom next year and a rebound soon after. Now it says "at least three more years of pain."
Why the deepening pessimism?
Three pressures stacking up: memory chip prices rising, pushing up bill-of-materials costs; CPU supply constraints slowing finished-product shipments; and the Windows 10 end-of-support upgrade wave fading.
This means → the first two squeeze the supply side, the third weakens demand — both ends deteriorating simultaneously explains three straight years of downward revisions.
Notably, average selling prices (ASPs) are actually climbing — spec upgrades and higher component costs are pushing per-unit prices up, so revenue declines (2026: down 5%) are far milder than shipment declines (down 14%).
Shipments are falling — why isn't revenue falling as fast?
Goldman forecasts PC revenue down 5% in 2026, down 2% in 2027, and up 3% in 2028.
In plain terms = fewer PCs sold, but each one sells for more — price increases partly offset the volume drop.
This reflects a structural shift from "selling on volume" to "selling on price." For upstream memory and CPU makers, shrinking volumes with rising ASPs mean profit resilience may be better than headline shipment data suggests.
What makes AI PCs and gaming PCs the outliers?
AI PCs: Goldman projects 150 million units shipped in 2026, rising to 199 million by 2028 — a 15% CAGR. Penetration climbs from 59% to 82%.
This means → by 2028, more than four out of every five PCs sold will carry on-device AI capability. AI PCs are shifting from "new category" to "default configuration."
Gaming PCs: 26 million units in 2026, 28 million by 2028, a 4% CAGR versus -2% for the overall PC market. Revenue CAGR hits 7%, driven by GPU upgrades and thinner designs pushing ASPs higher.
Who is driving AI PC demand?
Goldman highlights ByteDance's new AI applications — OpenClaw, Seedance, and others — as key catalysts for AI PC penetration gains in 2026.
In plain terms = hardware alone isn't enough. Users need killer apps that make their old machines feel slow — ByteDance's AI tools are playing exactly that role.
This reflects a shift in AI PC growth logic: from "chipmakers push specs" to "app makers pull demand." The latter is a more direct trigger for consumer upgrade intent.
What does this mean for investors?
The overall PC chain faces a diverging landscape of shrinking volumes but rising prices: pure-assembly players face the most pressure, while memory and CPU makers with pricing power may actually benefit from higher ASPs.
This means → whether AI PCs and gaming PCs can sustain structural growth against a shrinking total market is the key test of the PC supply-chain divergence thesis.
In plain terms = don't fixate on the headline shipment number. The real investment signal lies in the structure — who is raising prices, and who is gaining share.
Content is for reference only, not financial advice.