By Alper Yegin, CEO of LoRa Alliance
After years of slow progress in building a global Internet of Things, we now have billions of connected devices and worldwide networks ready to do far more than just gather status updates. Physical AI is the key to unlocking this potential, enabling new features and uses that can make IoT systems more independent and intelligent.
Global IoT coverage, expected to exceed 21.1 billion devices this year, has largely been built on four main wireless technologies: cellular, Wi-Fi, Bluetooth, and LoRaWAN. While connectivity has brought us this far, simply linking devices everywhere was never the ultimate goal for IoT.
What we’ve achieved so far is just the beginning of a much larger vision—making IoT even more widespread and vital, with diverse industrial applications designed to tackle some of the toughest challenges users face.
How can IoT reach this goal? How can connected devices have a more active, smart, and valuable impact on their systems and the world around them?
The answers lie in the fast-growing combination of IoT and AI, where IoT’s physical networks meet AI’s digital intelligence, sparking a new trend: physical AI.
What is physical AI?
There are various ways to define physical AI, but simply put, it involves AI systems that receive real-world data from sensors and can trigger actions in the physical world. If that sounds like a perfect fit for IoT, it is.
In 2026, physical AI is widely recognized. Major global companies like Nvidia, SpaceX, Tesla, Amazon, and Broadcom are investing billions in this area. For them, physical AI often relates to humanoid robots and self-driving vehicles.
These examples show a direct blend of the physical and digital—giving AI a physical form.
However, physical AI also has great potential for many current business and industrial IoT applications, and in some cases, it’s already being used.
Furthermore, there’s a growing chance to combine IoT and AI to create even more physical AI uses and bring advanced AI features to future connected devices.
How physical AI benefits IoT (and vice versa)
IoT and AI started working together long before physical AI became popular. Early AI was used to analyze digital data, often historical information stored in the cloud for reports. In those cases, AI’s role in IoT was limited to analysis.
With physical AI, however, AI now interacts with the physical world by getting real-time data from IoT sensors, processing it at the edge—on the device or a server—and starting immediate physical responses.
This is already happening. Real examples of physical AI are running in IoT networks globally. For instance, some IoT security cameras now use AI to analyze movement and send alerts.
For example, cameras in a remote forest can detect wildfires or specific animals and trigger warnings. Or, in a store, cameras with built-in AI can count customers and notify staff when a certain number is reached.
Other examples are in industrial settings. Vibration sensors on IoT-connected machines in places like factories or refineries can use on-device AI to predict equipment failures. Companies like Honeywell, Advantech, Watteco, and TE Connectivity already offer such products.
For IoT users, the real advantage is moving from simply knowing conditions to enabling AI-driven actions that add clear value. Physical AI can lead to better efficiency, cost savings, or even new revenue opportunities.
AI’s analysis of physical data can also guide IoT strategy. Insights can help decide where to add more sensors, meaning the next 21 billion IoT devices might be deployed with help from physical AI.
Also, the benefits go both ways. While IoT gains operational and financial advantages, IoT connections can improve AI by providing real-time, changing data from the physical world to large language models (LLMs) and chatbots.
The role of LoRaWAN in physical AI
As one of the four main wireless IoT technologies, LoRaWAN plays a key role in the physical-digital merge leading to the Physical AI era. In fact, LoRaWAN is in a unique position to help AI interact with the physical world. With over 125 million connected devices, it’s the most widely adopted LPWAN technology globally.
Its strength isn’t just in the number of devices, but also in supporting the widest range of applications among wireless IoT technologies. Wherever it’s used, LoRaWAN offers long-range, deep indoor coverage and low power use, allowing Physical AI to run almost independently and close to the data source, ensuring quick results.
In any remote or hard-to-reach location where an IoT sensor and its data are, LoRaWAN—available globally via terrestrial and satellite base stations—can likely reach it. This accessibility is supported by a strong ecosystem of low-cost, certified devices for various settings.
The shift to physical AI has just started, but LoRaWAN-connected devices are among the earliest testbeds for this technology. What happens over these connections, and how AI’s ability to interact with the physical world grows, will shape the next wave of IoT deployments and the value they offer users.
Author biography:
Alper Yegin is the CEO of the LoRa Alliance. He leads the organization’s strategy and supports the growth and global use of LoRaWAN, a key standard for low-power wide-area networks (LPWAN) in the Internet of Things (IoT). Before becoming CEO, he chaired the LoRa Alliance Technical Committee for eight years and served as Vice-Chair of the board for seven years. With over 25 years in IoT, mobile, and wireless communications, Yegin has held senior roles, including CTO at Actility, and positions at Samsung Electronics, DoCoMo, and Sun Microsystems. He has contributed to global standards in groups like IETF, 3GPP, ETSI, Zigbee Alliance, WiMAX Forum, and IPv6 Forum, where he also had leadership roles. Additionally, Yegin holds 16 patents and has written many technical standards and papers.
