Mobilint is a fabless AI chip company designing edge AI accelerators that target both large language models and vision workloads with tight power envelopes. In this video they walk through their Aries and Regulus product families, showing how an 80 TOPS NPU at around 25 W can be packaged as PCIe and MXM modules and even as a compact standalone AI box capable of running multi-billion-parameter LLMs entirely offline on a desk or at the edge. https://www.mobilint.com/aries/mlx-a1
—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—
The Aries NPU appears here as a PCIe accelerator and as an MXM module integrated into the MLX-A1 system, turning a small embedded PC into a powerful on-premises inference node for LLMs and computer vision. The demo includes a four-card configuration in a single chassis, effectively scaling to roughly 320 TOPS at about 100 W, which is a useful performance-per-watt point for edge servers, industrial PCs and compact AI gateways that need dense compute in a constrained platform.
Regulus targets even more constrained form factors as a full system-on-chip and system-on-module with integrated CPU, NPU and memory so that a complete mini PC class design can fit on a tiny board. With this module, Mobilint focuses on robots, drones, smart factory equipment and smart cars that must execute perception and control locally, without relying on cloud connectivity, enabling deterministic latency and privacy-preserving compute at the edge.
A key part of Mobilint’s pitch is their quantization and software stack: models are converted from 32-bit floating point to 8-bit integer with reported accuracy loss below one percent, which makes it realistic to replace very large foundation models by smaller, quantized variants while keeping most of the quality. In the demo they run an 8-billion-parameter LG LLM fully on the local accelerator and discuss how a well-quantized 7B-class model can deliver roughly 99% of the perceived performance of a model that might originally have hundreds of billions of parameters, all within the power budget of a compact edge device.
Recorded at Supercomputing SC25, the conversation also touches on architecture trends, with Mobilint arguing that future AI supercomputers and distributed systems will rely heavily on many ARM-based SoCs paired with NPUs rather than a small number of large GPUs. Their roadmap emphasizes edge servers, industrial desktops and embedded systems where dozens or hundreds of Aries and Regulus based boards can be deployed close to the data source, forming a scalable, power-efficient ecosystem.
I’m publishing about 90+ videos from Embedded World North America 2025, I upload about 4 videos per day at 5AM/11AM/5PM/11PM CET/EST. Join https://www.youtube.com/charbax/join for Early Access to all 90 videos (once they’re all queued in next few days) Check out all my Embedded World North America videos in my Embedded World playlist here: https://www.youtube.com/playlist?list=PL7xXqJFxvYvjgUpdNMBkGzEWU6YVxR8Ga
This video was filmed using the DJI Pocket 3 ($669 at https://amzn.to/4aMpKIC using the dual wireless DJI Mic 2 microphones with the DJI lapel microphone https://amzn.to/3XIj3l8 ), watch all my DJI Pocket 3 videos here https://www.youtube.com/playlist?list=PL7xXqJFxvYvhDlWIAxm_pR9dp7ArSkhKK
Click the “Super Thanks” button below the video to send a highlighted comment under the video! Brands I film are welcome to support my work in this way 😁
Check out my video with Daylight Computer about their revolutionary Sunlight Readable Transflective LCD Display for Healthy Learning: https://www.youtube.com/watch?v=U98RuxkFDYY
