Microsoft Technology Licensing LLC has published a patent application describing a system to leverage idle GPU capacity on edge devices such as gaming PCs, laptops and smartphones for micro-training of AI models.
The document, titled Opportunistic Micro-training on Edge Devices Using Spare GPU Utilization and numbered US 2026/0349820 A1, was published by the USPTO on November 5, 2026. It was filed on April 30, 2025, with inventor Aditya Vasanth Challapally.
How the system would work
According to the description, large training jobs would be split into independent micro-tasks sized to each device’s available capacity. The system would monitor GPU and CPU utilization, memory, temperature and power draw, assigning tasks only when spare capacity exists and backing off if the device heats up or becomes busy.
Participation would be opt-in. Users could start or stop tasks manually and set usage limits. Results would be cross-checked across devices to catch errors or manipulation.
Compensation and limitations
In the specification (paragraph [0019]), the document states the system “may compensate the owners of the edge devices” with credits redeemable for free software, access to the trained model or money. However, none of the application’s 20 claims includes mandatory payment or credit to owners.
Microsoft has announced no product, service, reward rates or launch date based on this filing. Patent applications describe concepts and do not guarantee commercial implementation.
Sources: USPTO (US 2026/0349820 A1), PCMag, PC Games Hardware.
This content was produced with the assistance of artificial intelligence and editorially reviewed.
Por Grok
