It’s been a couple of years since I wrote an article about different wireless technologies for IoT. To truly understand what’s happening in the air, we first need to look around and identify the trends emerging in the IoT space.
Edge, AI, and Space: what’s next for IoT networks?
Previous blogposts have touched upon topics like security and AI, which continue to grow as strong trends within IoT. AI thrives on data, and IoT has the capacity to generate immense quantities of it, making AI the perfect companion for data analytics. But beyond analytics, AI can also be a powerful tool to protect and secure IoT systems against threats and vulnerabilities. AI’s potential in IoT goes far beyond traditional data processing—it aims to integrate into IoT devices, enabling even simple units to make smarter, local decisions while also shaping how networks will need to evolve.
Edge Computing: handling data locally
Edge computing addresses the limitations of centralized cloud solutions by processing data locally, closer to the data source. This minimizes latency, which is critical for applications requiring uninterrupted operation and low response times. Think autonomous vehicles or industrial automation, where high reliability and resilience are essential. It’s also beneficial for energy efficiency, as processing closer to the source reduces the energy consumed in data transmission. On the downside, edge computing requires more energy than sensors that “merely” measure and transmit data, meaning the devices often cannot rely on battery power.
Edge computing is particularly suited for distributed AI systems requiring decentralized decisions. Yet despite its advantages, it has limitations. Edge devices often struggle to handle complex models that need continuous updates from centralized data sources. AI models are usually trained on aggregated data, which edge systems rarely have access to. Additionally, small edge devices lack advanced resources, like GPU acceleration, making computation-heavy processes less efficient than in a data center. The more processing you demand of edge devices, the higher the hardware cost. Therefore, local AI in simple IoT devices should currently be seen as an enhancement rather than a replacement for cloud-based AI.
It’s happening in the air
In 2025, many LPWAN technologies are preparing to move into space. Non-terrestrial networks (NTN), consisting of base stations orbiting Earth, will make it possible to cover every corner of the globe. Both LoRaWAN and 5G already function via satellite. Since 5G Release 17, satellites can serve as RF repeaters without decoding packets. In Release 19, not yet launched, satellites will gain the capability to receive and forward data packets. It is also likely that other LPWAN RF technologies, such as Mioty, will enter space by 2025. Even Starlink is set to launch IoT services via satellites in 2025, albeit using LTE CAT-1, CAT-1 Bis, and CAT-4 technologies, which are not LPWAN (i.e., not “Low Power”).
AI plays an even more significant role than previously imagined in defining what happens in the networks and how they need to evolve. I’ve observed technologies like 4G/5G and LoRaWAN seemingly moving in the same direction.
5G RedCap: a New LPWAN technology?
Staying earthbound for a moment, my forecast is that 5G RedCap (Reduced Capacity) will gain traction. RedCap is something of a middle ground—it’s not the fastest, nor the most energy-efficient, but it offers around 150 Mbps and is designed for slightly more demanding sensor applications. RedCap might become the technology capable of supporting the next generation of sensors with stable connectivity. It also works across frequencies used by private 5G networks, where current 3GPP LPWAN technologies (NB-IoT and LTE-M) do not operate.
RedCap delivers both low latency and high transfer speeds, making it suitable for scenarios where sensors must process large amounts of data locally, potentially using AI, and operate in slightly more challenging environments. In such cases, RedCap seems well-suited for the task.
LoRaWAN in 2025
As for LoRaWAN, in late 2024, Cisco announced its exit from the LoRaWAN market. Starting in 2025, it will no longer sell LoRaWAN products, and by 2030, support will cease entirely. Cisco has no planned migration path for existing customers, meaning organizations relying on its LoRaWAN solutions must seek alternative vendors to ensure continuity.
On a broader scale, the IoT market is trending towards greater encryption, easier software updates, and enabling local AI processing of data. However, these advancements push LoRaWAN away from being a pure LPWAN technology toward requiring more energy-intensive hardware.
The LoRa Alliance has published a roadmap outlining planned improvements to the LoRaWAN standard. The most notable is introducing “crypto agility,” allowing the use of both current and future encryption methods alongside LoRaWAN’s link layer, without requiring structural changes. This ensures LoRaWAN can adapt and implement new encryption methods as they emerge, keeping networks secure against evolving threats.
Through over-the-air updates (FOTA), devices can be safely updated, reducing vulnerability risks. These planned enhancements demonstrate the industry’s focus on strengthening security protocols and safeguarding networks from potential threats in 2025 and beyond. Discussions are also underway about integrating edge computing into the LoRaWAN architecture to improve network performance and efficiency, enabling faster data management and decision-making at the data source.
Open RAN: a more open network for IoT’s future
One technology I’m keeping an eye on is Open RAN—a network architecture that departs from the traditional proprietary standards where hardware and software are locked to one vendor. Instead, Open RAN leverages standardized interfaces and components, allowing operators to mix and match hardware and software from different providers. This approach aims to reduce costs, increase flexibility, and accelerate innovation in mobile networks.
Driven by organizations like the O-RAN Alliance, with support from tech giants like Nokia, Ericsson, and Qualcomm, Open RAN is expected to gain significance between 2025–2030 as more network operators implement it for both 5G and future IoT solutions. Early pilots and implementations are already emerging, particularly in regions where cost-effective solutions are crucial for expanding network coverage.
For IoT, Open RAN offers the opportunity to create networks where technologies like LoRaWAN and 5G not only coexist but integrate seamlessly. This paves the way for new applications, particularly in environments where diverse technologies must work together—from industrial IoT systems to smart cities.
This was a snapshot of what’s happening in connectivity for 2025. What do you think? What do you believe will be most important to watch in 2025?