On October 29, VeriSilicon (688521.SH) announced the launch of GLTiny, an innovative 3D graphics processing (GPU) programming model designed specifically for wearable devices with constrained power and memory resources. GLTiny provides a streamlined application programming interface (API) paired with an ultra-lightweight driver stack, enabling developers to harness the 3D graphics capabilities of VeriSilicon’s Nano OpenGL ES GPU IP family and deliver rich 3D visual experiences for next-generation smartwatches, smart glasses, and other resource-constrained wearables.
In conventional wisdom, 3D rendering is a heavy consumer of power and memory, inherently at odds with wearables that operate on milliwatt-scale power budgets and megabyte-scale memory footprints. GLTiny’s approach is to trim the graphics stack to the bone: an optimized programming model fully exploits the 3D processing capability of OpenGL ES GPUs to achieve efficient 3D graphics under extremely low power and limited memory, so that animated watch faces, 3D map navigation, and health-data visualization no longer depend on an over-provisioned application processor, but are instead rendered on demand by a dedicated GPU IP. This extends VeriSilicon’s earlier low-power video and audio DSP IP playbook — replacing over-spec general-purpose computing with purpose-built hardware that is just enough.
From an industry-landscape perspective, GLTiny’s arrival reflects a philosophical shift in hardware design for the on-device AI era: as smart glasses and watches become new entry points for AI interaction, both graphics rendering and AI inference must fit within strict power budgets, and the combination of lean interfaces, dedicated accelerators, and lightweight software stacks is replacing brute-force core counts. For China’s IP industry, VeriSilicon’s pairing of the Nano GPU IP family with a companion programming model delivers hardware and software as one package — upgrading IP licensing from turnkey circuits to a developer ecosystem. It is also a signal that domestic IP is reaching for incremental growth in high-barrier niches such as wearables.
Junyi Microelectronics likewise adheres to the design philosophy of putting just enough computing power in the right place: the J-3568 core board serves human-machine interaction terminals with a mature process and a balanced GPU/NPU configuration, while the K-series edge AI motherboards integrate stronger graphics and AI capability within a limited power envelope, powering smooth HMI rendering and multimodal interaction on touch-panel all-in-one machines. In future screen-as-entry-point scenarios such as smart cockpits and industrial handheld terminals, Junyi’s board platforms can work in concert with lightweight graphics stacks like GLTiny to bring 3D visualization ever closer to the field.
Source: https://www.yicai.com/news/103414199.html (October 29, 2026)