STMicroelectronics held a virtual media briefing, hosted by Patrick Aidoune, General Manager, General Purpose MCU Division at ST, on November 17, 2025. The briefing was held before their flagship event, the STM32 Summit, where they launched STM32V8, a new generation of STM32 microcontrollers.
STMicroelectronics introduced its new generation microcontroller, STM32V8, under the STM32 class recently. Built on an innovative 18nm process technology with FD-SOI and phase change memory (PCM) technology included, this microcontroller is the first of its kind in the world. It is the first under 20nm process to use FD-SOI along with an embedded PCM technology.
FD-SOI Technology
The FD-SOI is a silicon technology, co-developed by ST, which brought innovation in the aerospace and automotive applications. The 18nm process, co-developed with the Samsung Foundry, provides a cost-competitive leap in both performance as well as power consumption.
The FD-SOI technology gives a strong robustness to ionising particles and reliability in harsh operating environments, which makes it particularly suitable for intense radiation exposure found in earth orbit systems. The FD-SOI also helps reduce the static power consumption, along with allowing operations on a lower voltage supply, while sustaining harsh industrial environments as well.
Key Features
STM32V8’s Arm Cortex-M85 core, along with the 18nm process, gives it a clock speed of up to 800MHz, making it the most powerful STM32 ever shipped. It has also been embedded with up to 4 Mbytes of user memory in a competitive dual bank, allowing bank swapping for seamless code updates.
Keeping in mind the needs of developers, the STM32V8 provides for more compute headroom, along with more security and improved efficiency. Compared it is 40nm process node with the same technologies, the STM32V8 brings with it improved performance, higher density, and better power efficiency.
Industrial Applications
This new microcontroller is a multipurpose system to benefit several industries:
- Factory Automation and Robotics
- Audio Applications
- Smart Cities and Buildings
- Energy Management Systems
- Healthcare and Biosensing
- Transportation (ebikes)
Achievements
ST’s new microcontroller has been selected by SpaceX for its high-speed connectivity system in the Starlink Satellite System.
“The successful deployment of the Starlink mini laser system in space, which uses ST’s STM32V8 microcontroller, marks a significant milestone in advancing high-speed connectivity across the Starlink network. The STM32V8’s high computing performance and integration of large embedded memory and digital features were critical in meeting our demanding real-time processing requirements, while providing a higher level of reliability and robustness to the Low Earth Orbit environment, thanks to the 18nm FD-SOI technology. We look forward to integrating the STM32V8 into other products and leveraging its capabilities for next-generation advanced applications,” said Michael Nicolls, Vice President, Starlink Engineering at SpaceX.
STM32V8, like its predecessors, is expected to draw significant benefit from ST’s edge AI ecosystem, which is under continued expansion. Currently, the STM32V8 is in early-stage access for selected customers with key OEMs’ availability as of the first quarter 2026 and with broader availability to follow.
Apart from unveiling the new generation microcontroller, ST also announced the expansion of its STM32 AI Model Zoo, which is part of the comprehensive ST Edge AI Suite of tools. The STM32 AI Model Zoo has more than 140 models from 60 model families for vision, audio, and sensing AI applications at the edge, making it the largest MCU-optimised library of its kind.
This AI Model Zoo has been designed, keeping in mind the requirements of both data scientists and embedded systems engineers, a model that’s accurate enough to be useful and that also fits within their energy and memory constraints.
The STM32 AI Model Zoo is the richest in the industry, for it not only offers multiple models, but also scripts to easily retrain models, evaluate accuracy, and deploy on boards. ST has also introduced native support for PyTorch models. This complements their existing support for TensorFlow, Keras AI frameworks, LiteRT, and ONNX formats, giving developers additional flexibility in their development workflow. They are also introducing more than 30 new families of models, which can use the same deployment pipeline. Many of these models have already been quantised and pruned, meaning that they offer significant memory size and inference time optimisations while preserving accuracy.
Additionally, they also announced the release of STM32 Sidekick, their new AI agent on the ST Community, available 24/7. This new AI agent is trained on official STM32 documentation (datasheets, reference manuals, user manuals, application notes, wiki entries, and community knowledge base articles) to help users locate relevant technical data, obtain concise summaries of complex topics, and discover insights and documents. Alongside, they announced STM32WL3R, a version of their STM32WL3 tailored for remote control applications supporting the 315 MHz band. The STM32WL3R is a sub-GHz wireless microcontroller with an ultra-low-power radio.
~ Shreya Bansal, Sub-Editor
