IN Brief:
- Avnet Silica has raised its 2026 served-semiconductor-market forecast from $433 billion to $454 billion.
- DDR4 and LPDDR4 are under allocation, with DRAM lead times above 26 weeks and some programmable logic approaching 52 weeks.
- Industrial applications now represent 27% of the EMEA market tracked by Avnet Silica, valued at $12 billion.
Avnet Silica has raised its 2026 served-semiconductor-market forecast to $454 billion as strengthening demand begins to coincide with longer lead times and tighter availability across several component classes. The distributor’s Q3 Trendliner previously put the same market at $433 billion, while its latest forecast projects further growth to $503 billion in 2027 and $584 billion by 2029.
The figures exclude DRAM, flash, MPU compute, GPUs, and AI processors, concentrating instead on much of the analogue, logic, discrete, power, sensing, control, and embedded component base used across mainstream electronic systems. That makes the report particularly relevant to industrial and embedded designs, where improving overall semiconductor revenue does not necessarily translate into easier component sourcing.
European manufacturing indicators have improved alongside the market forecast. The Eurozone manufacturing output index remained in expansion for a fifth consecutive month in June, increasing from 51.3 to 51.7 and contributing to the sector’s strongest quarter since 2022. Export demand remains weak, however, while inventories continue to be reduced, leaving the regional recovery uneven.
The EMEA application mix is also changing. Automotive remains the largest semiconductor vertical tracked by Avnet Silica in the region at $17 billion, but industrial applications have increased their share from 25% to 27% and are now valued at $12 billion. Industrial automation is forecast to deliver a three-year compound annual growth rate of 10.8%, behind automotive high-performance computing at 21.5% and energy management at 13%.
Those growth areas increasingly draw on overlapping component families. Industrial automation, vehicle electronics, energy systems, edge computing, and communications can all require microcontrollers, sensors, programmable logic, memory, and power-management devices, creating competition for manufacturing capacity even where individual end markets are following different cycles.
Memory is currently showing some of the clearest constraints. DDR4 and LPDDR4 are under allocation, while Avnet Silica puts DRAM lead times above 26 weeks. Lead times have also increased for NAND flash, eMMC, and solid-state drives as demand, lifecycle decisions, and changing manufacturing priorities affect supply.
High-bandwidth memory used in AI infrastructure is one part of that pressure because it influences how manufacturers allocate advanced DRAM capacity, although the effect is not uniform across every memory technology. Established products can become harder to source at the same time that suppliers direct investment and production towards higher-growth device families.
Conditions are tightening elsewhere as well. Avnet Silica reports rising prices and lead times across some microprocessors and microcontrollers, limited capacity for selected sensor products, and lead times approaching 52 weeks for many programmable-logic devices. For products with lengthy qualification cycles, that can turn an apparently ordinary component choice into a programme constraint.
Industrial and embedded equipment often remains in production for substantially longer than consumer hardware, making redesign less attractive once software, certification, EMC performance, thermal behaviour, and manufacturing processes have been established around a particular device. Replacing an FPGA, MCU, sensor, or memory component can therefore involve considerably more work than changing a purchasing record.
Availability increasingly has to be considered alongside electrical performance during architecture development. Pin compatibility, package availability, software portability, alternative suppliers, lifecycle status, and qualification effort all affect how exposed a design becomes when lead times extend or allocation returns.
The stronger semiconductor outlook consequently carries a less convenient procurement picture. Avnet Silica’s Q3 figures show expanding demand across automotive, energy management, and industrial automation while memory, programmable logic, processors, microcontrollers, and selected sensors face varying degrees of constraint. Market growth and predictable supply are again proving to be separate conditions.
