Global smartphone shipments fell 20 per cent year on year in the third quarter of 2026, according to Counterpoint Research — the weakest third quarter in fourteen years.
The cause is not consumer demand. It is a shortage of memory chips, and the reason those chips are scarce is that the AI build-out is buying them.
What is actually happening
Investment in AI infrastructure has driven demand for memory chips up sharply. Those are the same components a phone needs, and AI projects consume them in volumes a handset line cannot outbid. Supply tightened and prices rose, so manufacturing costs rose with them.
Phone makers responded in two ways. Some are raising prices. Others have cut production and shipments deliberately, choosing to protect margin rather than chase volume — which is why a shortage shows up as fewer phones shipped rather than only as dearer ones.
Who is where
Samsung shipped more than anyone, holding 23 per cent of the market, helped by interest in the Galaxy Z Fold 8 series and a steadier supply of the parts mid-range phones need. Even so its own shipments were down 3 per cent on the same quarter last year.
Apple holds 22 per cent. Last year the full iPhone 17 line arrived together; this year the standard iPhone 18 model did not launch, which changed what reached retailers in the first month of the new range.
The connection most coverage will miss
This is the same shortage, measured at a different point in the chain.
In December we reported that RAM prices in Dhaka had nearly doubled since September, with sellers saying the problem was supply rather than demand. They were describing the retail end of exactly this: AI data centres absorbing memory, DDR4 getting scarce, a 16GB kit reaching Tk 10,000.
What Counterpoint adds is the scale. A PC buyer in Dhaka paying double for memory and a global shipment figure at a fourteen-year low are the same event seen from opposite ends, and the chain runs from a data centre procurement contract to a shop in Elephant Road.
What it means in Bangladesh
Three consequences follow, and the third is the one worth planning around.
Prices here will lag the global move rather than track it. Stock already in the country was bought at the old cost, so the increase arrives when that stock is replaced — which means the rise is still ahead, not behind.
The squeeze will land hardest on mid-range phones. A flagship can absorb a higher bill of materials inside a large margin; a Tk 20,000 handset cannot, and that is the segment most Bangladeshi buyers actually shop in.
And the refurbished market will grow, because people hold phones longer when new ones get dearer. That is already visible, and it raises a practical problem this site has covered before: a second-hand handset bought outside official channels carries the IMEI registration question, which the piece on the grey-market price gap sets out. A shortage that pushes buyers toward used phones pushes them toward that risk at the same time.
It also complicates the local assembly picture. Production here fell 24 per cent in a year on tax and grey-market pressure; a global component squeeze is a second force in the same direction, and it is one no domestic policy can relieve.
Source: Counterpoint Research, via Prothom Alo.
