Ericsson and Altair Semiconductor have completed what they describe as the first network connection using enhanced Reduced Capability, or eRedCap. The November 19 announcement described a test pairing Ericsson’s radio access network software with a device built around Altair’s ALT1550 modem. It is a step toward standalone 5G for wearables, trackers and sensors that do not need a phone’s data rates and cannot easily accommodate a phone-sized battery.
A Smaller Radio for a Different Job
Conventional 5G hardware is built to cover demanding workloads, including large downloads and mobile video. An asset tracker might send location updates intermittently, while a wearable assistant might transmit a short voice query or receive a response. The performance that makes a broadband device versatile can be an unnecessary design burden for products that move much less data.
Early 5G Reduced Capability, or RedCap, gave device designers a way to use standalone networks without implementing the full set of broadband capabilities. eRedCap, introduced in 3GPP Release 18, goes further toward the low-data-rate end of that range. Ericsson’s technical overview of RedCap distinguishes the kinds of workloads these technologies can support and describes eRedCap’s potential to lower device cost and energy use. That is a design goal, not a measured battery-life result from the November test.
Altair already makes LTE-M and NB-IoT chipsets for products such as smart meters and trackers. Its cellular IoT portfolio positions the ALT1550 as a way for customers building around those earlier connectivity options to consider standalone 5G. The test matters because a modem and a radio network have to work together before manufacturers and carriers can validate an actual product. It does not establish that every existing device can move to 5G without a new hardware design.
Where the AI Actually Runs
The modem supplies connectivity rather than inference. An AI device can run some processing locally and use the network when it needs outside information or a model running in the cloud. The modem test did not demonstrate either of those AI workloads, and the companies did not name a wearable ready to ship.
That distinction matters when deciding where eRedCap might fit. Glasses that handle wake-word detection locally and send occasional voice requests have different network needs from glasses that continuously upload video for remote interpretation. The former design can make a modest data link useful; the latter needs enough bandwidth and responsiveness to handle a sustained stream. In both cases, the battery also powers the camera, processor, microphones and any display. A simpler modem helps address one part of that budget without solving the rest.
What Still Has to Happen Before Rollout
Altair says the milestone keeps commercial deployments on track for 2027. That is a target, not a launch date or a promise of carrier coverage. The announcement does not identify participating carriers, device makers or markets, nor does it publish measured throughput, latency or energy consumption for the connection. The next evidence to watch is testing on commercial networks and devices that show what these trade-offs mean outside the demonstration.
For AI hardware makers, the immediate result is narrower but useful: a modem built for modest data rates has connected to Ericsson’s network software. A wearable with a focused set of AI features may not need a premium radio, but it still needs a network that supports the device and a product design that makes good use of the connectivity.