While many in the industry are focused on AI chip innovations for nearly 12,000 data centers, chips powering world’s 20+ billion IoT devices are undergoing significant innovations as well
Dec. 19, 2025 –
Prediction 1: Edge AI integration into IoT chips to accelerate
Prediction 2: The share of chiplet-based and RISC-V-based IoT chips to increase
Prediction 3: More IoT chips to be designed with carbon awareness in mind
Prediction 4: More IoT devices to be produced locally
Prediction 5: IoT chip design to become heavily AI-supported
Prediction 6. IoT security-by-design to become non-negotiable
While many in the semiconductor industry are focused on AI chip innovations for the world’s nearly 12,000 data centers, the chips powering the world’s 20+ billion IoT devices are undergoing significant innovations as well.
IoT semiconductors, or, specialized electronic components that enable the functionality and connectivity of IoT devices.
IoT devices, or, physical objects with embedded compute and network connectivity that can autonomously transmit or receive data without real-time human intervention. Typical devices include end devices and gateways such as smart meters, asset trackers, wearables, industrial sensors, building controllers, or smart home appliances.
This also includes connected automotive modules such as telematics units when acting as IoT endpoints or gateways. Excludes smartphones, tablets, PCs, infotainment systems, and automotive designs that do not act as IoT endpoints or gateways. It excludes devices with passive or non-networked connectivity, such as RFID tags or QR code scanners, and devices connected only within closed local networks.
IoT semiconductor functionality, or, any semiconductor component that has the main purpose to sense/actuate, compute, connect, manage power, or secure.
IoT semiconductors market 2024-2030
Prediction 1: Edge AI integration into IoT chips to accelerate
Edge AI integration into IoT devices will begin a major shift toward AI-capable hardware.
Most IoT devices today lack the built-in compute needed to run AI workloads. Even though demand for local inference has been rising to improve latency, resiliency, bandwidth efficiency, and privacy, the majority of today’s 21 billion deployed IoT endpoints still rely on external processing or simple rule-based logic. This gap between demand and capability sets the stage for a shift in 2026.
NPUs and AI-capable cores entering mainstream IoT designs
Vendors are expanding edge AI across IoT tiers. In recent years, only a small subset of IoT products (typically industrial gateways, advanced cameras, and high-end modules) have integrated NPUs or low-power AI accelerators. Vendors are now starting to push these capabilities into broader device categories. New, IoT SoCs are being designed with lightweight NPUs, vector extensions, and DSP-like AI cores to support tasks such as anomaly detection, small-model vision, local audio intelligence, and condition monitoring directly on the device.
Prediction: IoT Analytics expects 2026 to mark the first broad wave of IoT devices embedded with edge AI acceleration. Shipments of AI-enabled chipsets will expand into sensors, IoT connectivity modules, industrial PCs, and mid-tier gateways that previously lacked any on-device AI inference.
More complex SoC designs driving demand for AI-ready tooling
AI features affecting IoT chip design priorities. Embedding NPUs and AI blocks into IoT silicon has increased design complexity, especially around thermal budgets, verification, memory bandwidth, and security. As a result, IoT chip vendors are leaning more heavily on EDA tools optimized for AI compute analysis, reusable IP such as low-power NPUs and secure enclaves, and mature-node foundry processes tuned for mixed workloads (compute + connectivity + security). These needs are emerging across consumer, industrial, automotive, and energy IoT segments.
Prediction: IoT Analytics expects 2026 to bring wider adoption of AI-aware EDA flows and off-the-shelf AI IP subsystems in IoT chip development. These tools and IP blocks will reduce design complexity and lower the barrier for adding small-model inference to mass-market IoT devices.
Edge AI becoming a defining differentiator for IoT OEMs
Device makers linking AI to feature innovation. As AI-capable hardware becomes more accessible, device makers are beginning to treat local inference as a competitive discriminator, enabling features such as privacy-preserving analytics in smart home devices, real-time defect detection in industrial sensors, or offline wake-word detection in consumer electronics.