Samsung is cutting phone production by up to 30% in Q4 2026 — not because demand collapsed, but because memory costs have destroyed unit economics.
MoneyToday reports the Mobile eXperience division has already asked suppliers to reduce component deliveries by 20–30%. If that holds, full-year output could drop from 270 million units to the low 200s.
The culprit: 12GB LPDDR5X modules hit $145–$146 in Q2, up 175% year-over-year. Q3 saw another 20% spike, pushing the module toward $180. Industry sources say those prices wiped out margins on many models.
So Samsung is choosing volume discipline over margin erosion. They're refusing to build phones that lose money once the memory bill is in.
The irony: Samsung makes the handsets and produces the memory. But chipmakers — including Samsung — have shifted capacity toward high-bandwidth memory for AI data centers. That squeezed supply of conventional DRAM phones actually need.
One division wins. The other gets crushed.
Cutting phone output protects the handset P&L. It doesn't bring conventional DRAM capacity back online. The alternative — raising handset prices high enough to restore margin — risks killing demand further.
Samsung is betting that unsold, margin-poor phones are worse than empty shelf space.
This isn't a seasonal slowdown. Seasonal pauses don't trigger 20–30% supplier cuts. The scale points to memory economics as the decisive factor, not the calendar.
And Samsung's rivals face the same memory bill. Executives have already framed the shortage as an industry problem. Anyone still planning a full Q4 build using last year's memory math is using a cost base that no longer exists.
One caveat: Samsung hasn't confirmed the figures. Supplier requests aren't shipments. A 20% cut and a 30% cut are not the same year. Final production can still shift.
But the signal is clear: when your memory costs more than your phone, volume is the lever — not price.
MoneyToday reports the Mobile eXperience division has already asked suppliers to reduce component deliveries by 20–30%. If that holds, full-year output could drop from 270 million units to the low 200s.
The culprit: 12GB LPDDR5X modules hit $145–$146 in Q2, up 175% year-over-year. Q3 saw another 20% spike, pushing the module toward $180. Industry sources say those prices wiped out margins on many models.
So Samsung is choosing volume discipline over margin erosion. They're refusing to build phones that lose money once the memory bill is in.
The irony: Samsung makes the handsets and produces the memory. But chipmakers — including Samsung — have shifted capacity toward high-bandwidth memory for AI data centers. That squeezed supply of conventional DRAM phones actually need.
One division wins. The other gets crushed.
Cutting phone output protects the handset P&L. It doesn't bring conventional DRAM capacity back online. The alternative — raising handset prices high enough to restore margin — risks killing demand further.
Samsung is betting that unsold, margin-poor phones are worse than empty shelf space.
This isn't a seasonal slowdown. Seasonal pauses don't trigger 20–30% supplier cuts. The scale points to memory economics as the decisive factor, not the calendar.
And Samsung's rivals face the same memory bill. Executives have already framed the shortage as an industry problem. Anyone still planning a full Q4 build using last year's memory math is using a cost base that no longer exists.
One caveat: Samsung hasn't confirmed the figures. Supplier requests aren't shipments. A 20% cut and a 30% cut are not the same year. Final production can still shift.
But the signal is clear: when your memory costs more than your phone, volume is the lever — not price.