IoT module shipments equipped with artificial intelligence fell by 17% year-on-year in the first quarter of 2026, according to data from Counterpoint Research. This decline ended a period of steady growth recorded throughout 2025.
The research firm attributed this downward trend to the rising cost of memory components. Consequently, these expenses had a greater impact on units that require high on-board processing capabilities.
Impact of Memory Costs on Hardware
Hardware configurations played a significant role in the market shift. Specifically, modules fitted with central processing units (CPUs), graphics processing units (GPUs), and neural processing units (NPUs) for localized computing drove the overall decline.
“The different cost structure and component dependencies affected the growth and adoption of these two approaches.”
Tina Lu, Senior Analyst at Counterpoint Research
Alternative Designs in IoT Module Shipments
Meanwhile, modem-AI modules, which integrate intelligence directly into the modem, recorded a modest increase. Notably, these specific variants do not rely on external memory and do not perform application processing.
This alternative design has found traction in specific sectors. As a result, the demand for these modules remains steady in enterprise environments.
Market Drivers and Price Increases
Mohit Agrawal, director of the IoT practice at Counterpoint Research, linked the positive performance of modem-based modules to specific applications. Specifically, these include routers used for 5G fixed wireless access alongside various enterprise deployments.
However, the average selling price of AI-embedded modules experienced a double-digit increase. This rise, driven by a more expensive bill of materials, ultimately reduced overall customer demand.
Overall Market Outlook
Despite the drop, the broader cellular market remains largely unaffected by these shifts. Currently, AI-embedded units represent only 6% of total IoT module shipments.
Therefore, analysts expect the wider industry to maintain stability. The sector continues to monitor how memory pricing affects IoT module shipments across global markets.
Analysis of Component Dependencies
The distinction between CPU-driven modules and modem-based modules highlights a critical shift in hardware design. Specifically, developers must choose between localized processing power and cost efficiency. Furthermore, the rising bill of materials makes high-end modules less attractive for budget-conscious enterprise deployments.
In addition, the demand for 5G fixed wireless access routers continues to support the modem-AI segment. Consequently, this specific application provides a steady stream of orders for manufacturers. Meanwhile, the broader cellular market remains resilient despite these localized fluctuations.
Future Outlook for Cellular IoT
Looking ahead, the market share of AI-embedded units is expected to remain a minor portion of the overall cellular segment. Specifically, the 6% share recorded in the first quarter suggests that standard IoT modules still dominate the industry. Therefore, the overall impact of memory pricing on the broader market will likely remain limited.
However, manufacturers must address the double-digit rise in average selling prices to stimulate future demand. As a result, the industry may see a shift toward more cost-effective designs. Ultimately, the balance between processing power and component costs will shape the next phase of cellular hardware.
