Demand for memory used in artificial-intelligence infrastructure is pushing manufacturers away from sub-$100 phones, shifting part of the cost of the AI boom onto consumers who may never use an AI model.
Shipments of smartphones costing less than $100 fell 34% across Africa in the second quarter of 2026, as rising memory prices forced manufacturers to raise prices, cut specifications or stop producing some low-cost devices. Royal Ibeh of BusinessDay reported Thursday that growing demand for memory chips from the artificial-intelligence industry is tightening component supplies at the same time African consumers remain heavily dependent on inexpensive handsets for internet access.
A smartphone is often the primary computer for consumers across the continent. Losing the cheapest devices therefore affects access to banking, education, government services, messaging, commerce and employment rather than simply changing which phone someone buys.
Manufacturers build entry-level phones by controlling every component cost closely. Memory and storage represent a meaningful share of that bill. When chip suppliers can earn more serving data centers and high-end computing markets, low-margin handset manufacturers have less room to compete for the same production capacity.
AI investment is changing who gets priority.
Technology companies building large computing clusters can spend billions of dollars on servers and memory. A manufacturer selling a smartphone for $80 or $90 operates under a completely different margin structure. Even modest increases in component prices can force the company to charge more or remove the device from the market.
The scarcity therefore redistributes access through purchasing power. Wealthier technology firms can pay the higher price of components because computing capacity supports products expected to generate large returns. Low-income consumers encounter the same increase as a barrier to purchasing their first smartphone.
Africa remains particularly exposed because smartphone ownership is still far from universal. BusinessDay reports that roughly 76% of people in Sub-Saharan Africa do not own a smartphone, leaving a large population whose path onto the mobile internet depends on prices continuing to fall.
The development complicates a long-running assumption about technology diffusion. Advanced technology typically becomes cheaper over time, allowing products first purchased by wealthy consumers to reach broader markets. Smartphone adoption followed that pattern as manufacturers created increasingly capable devices below $100.
AI infrastructure can temporarily reverse the process when it competes for the same physical inputs.
The effect also exposes how global technology markets allocate scarce capacity. Semiconductor companies do not produce memory according to social need. They direct investment toward customers and products likely to generate the strongest returns. When AI infrastructure offers better margins, production follows the money.
African governments have limited control over that allocation because most of the relevant semiconductor manufacturing occurs elsewhere. Policies intended to expand digital inclusion can subsidize connectivity or lower taxes on devices, but they cannot easily solve a global shortage of components.
Phone manufacturers may respond by using less memory, extending the life of older components or moving entry-level devices above the $100 threshold. Each option changes the experience for consumers. Cheaper phones can become slower or less capable, while higher prices delay adoption altogether.
The AI boom is usually measured through data centers, venture funding and corporate productivity. Component markets reveal another accounting. Computing resources flowing toward the richest technology companies can increase the cost of basic digital access somewhere else.
If memory demand continues rising, Africa’s smartphone market may divide more sharply between consumers able to move into higher-priced devices and people who remain on feature phones or outside mobile internet access entirely. AI could expand what computers are capable of while simultaneously increasing the price of owning the computer millions of people are still trying to obtain.