Always-on AI meets the realities of skin-contact thermals – and the need for a new micro-cooling architecture
Every computing platform eventually encounters a physical boundary. For AI glasses, that boundary is heat – and it sits directly on the human face.
The category’s ambition is clear: always-on assistants, real-time translation, contextual awareness layered seamlessly onto the physical world. Unlike smartphones, which are used in bursts and then set aside, AI glasses are designed for continuous, all-day operation. That shift fundamentally changes the engineering problem. Compute must be sustained over time, and the resulting heat must be managed within a device that rests on skin.
In this form factor, performance is shaped as much by human tolerance as by silicon capability.
The Skin-Contact Constraint
Wearables operate under a different set of thermal rules than any other consumer electronics category. Devices like smartwatches, fitness bands, and wireless earbuds all share a defining characteristic: they maintain prolonged contact with the body. AI glasses take that constraint further. They sit on some of the most thermally sensitive areas of the body – the nose bridge and temples – where even small increases in temperature are immediately noticeable.
International safety standards reflect this sensitivity, capping surface temperatures for skin-contact devices at approximately 48°C. In practice, product teams target closer to 41-42°C to ensure comfort over extended wear. Within that narrow thermal window, every watt of power and every degree of temperature rise must be carefully managed.
In AI glasses, this constraint has become a central design variable. The frame offers minimal surface area to distribute heat and almost no internal volume to buffer it. At the same time, the device is expected to operate continuously, often in direct sunlight or warm environments, with no opportunity to cool down between tasks. Heat accumulates quickly, and once the frame approaches its comfort threshold, the system must respond – scaling back performance, shortening usage windows, or limiting features.
This is already visible in current devices. Smart glasses can restrict high-resolution video capture to short durations. As cameras, processors, and sensors engage simultaneously, the system approaches its thermal ceiling and adjusts in real time to stay within safe and comfortable limits. The experience is governed by heat as much as by software.
Why AI Changes the Equation
For years, passive cooling strategies were sufficient for wearable devices. Heat spreaders, graphite layers, and thin metal frames distributed thermal energy across the device and released it gradually into the surrounding air. This approach aligned with intermittent workloads and modest power levels. But AI introduces a different operating profile.
On-device inference, real-time computer vision, and continuous sensing keep processors active for extended periods. Instead of short bursts followed by idle time, the system operates closer to a steady state. Heat builds and persists.
In AI glasses, the implications are immediate. The form factor provides limited surface area for heat spreading and minimal mass to absorb thermal spikes. Weight constraints further restrict the use of larger passive components, as every gram affects comfort and wearability. As a result, the system reaches its thermal limits faster and remains near them during sustained use.
Why Spreading Heat Reaches Its Limits
Most thermal solutions in compact electronics are built around conduction – moving heat away from the source and distributing it across a larger area. Materials like graphite sheets and vapor chambers excel at this task.
But once heat reaches the surface of the device, another process takes over: convection. In still air, a thin layer of warm air forms around the frame, slowing the transfer of heat into the environment. This thermal boundary layer becomes a bottleneck, limiting how quickly heat can leave the system.
At that point, additional improvements in conduction provide diminishing returns. The heat has already been spread; the challenge is removing it efficiently from the surface.
In AI glasses, where airflow is minimal and space is constrained, this transition happens quickly. The system becomes limited by how effectively it can move heat away from the frame, not just within it.
A Frame with Multiple Heat Sources
The processor is only one contributor to the thermal load.
As AI glasses evolve, they integrate additional components that generate heat continuously. Cameras, sensor arrays, wireless radios, and – critically – display engines for augmented reality all operate within the same confined structure. In AR glasses, light engines project images into the lens, introducing another sustained heat source within millimeters of the processor.
These elements interact thermally. As internal temperatures rise, performance can shift – frame rates drop, image quality changes, and system responsiveness declines. At the same time, the external frame temperature approaches the user’s comfort threshold.
Managing this system requires balancing internal performance with external wearability, both governed by the same thermal envelope.
Airflow, Engineered for the Micro-Scale of Eyewear
Improving thermal performance in AI glasses requires more than redistributing heat – it requires removing it efficiently from the system. That introduces airflow into a form factor that has historically had none.
Traditional fans are not suited to eyewear. They require space, generate audible noise, and rely on mechanical components that introduce wear over time. These constraints have limited their use to larger devices like laptops.
Advances in solid-state micro-cooling, like those achieved by xMEMS, offer a different approach. By generating directed airflow at the millimeter scale, these systems introduce convection exactly where it is needed – at the heat source and along the thermal path through the frame.
Within the temple arm of a pair of glasses, this can be implemented as a controlled airflow channel. Air enters through a small inlet, passes over heat-generating components, and exits through an outlet at the end of the frame. The flow is continuous, silent, and precisely directed.
What makes this effective is placement. Even modest airflow, applied at the right location, can significantly reduce both internal component temperatures and the temperature felt at the surface.
From Short-Duration Features to Sustained, All-Day Usage
Thermal design determines how AI glasses are experienced in practice.
Today, many features are constrained by thermal limits – high-resolution video recording, continuous AI processing, and extended camera use are often time-bound or dynamically adjusted. These constraints shape how and when the device can be used.
As thermal headroom increases, those boundaries expand. Features that were previously limited to short sessions can operate for longer durations. AI assistants can remain active throughout the day. Real-time processing becomes more consistent and predictable.
This progression follows a familiar pattern in computing. Removing a physical constraint – whether in power delivery, memory bandwidth, or storage – enables new usage models and expands the role of the device.
AI glasses are approaching that same transition, with thermal management as the gating factor.
Engineering for Continuous Wear and Continuous Compute
The success of AI glasses depends on delivering sustained performance within a form factor designed for comfort. That balance sits squarely in the thermal architecture.
As compute density increases and new capabilities are added, the amount of heat generated within the frame will continue to rise. Addressing this requires a combination of approaches: conductive materials to distribute heat, paired with targeted airflow to carry it away from the system.
Micro-cooling complements existing thermal strategies by introducing controlled heat removal at a scale compatible with eyewear. It adds a capability that passive solutions alone cannot provide, enabling a more balanced and resilient thermal system.
As AI glasses evolve from early products into mass-market devices, thermal design will play a central role in shaping both performance and wearability.
Mike Housholder is VP and GM of xMEMS’ Thermal Management Business Unit.