SiMa.ai has released Agentic, a development platform specialized for the development of Physical AI applications. The product focuses on reducing complex development processes through natural language interfaces and lowering the entry barriers of existing GPU-centric development environments.

Syma AI announced on the 17th that it has launched 'Palette Neat,' an agentic development platform for physical AI. The company explained that this platform reduces the development time for complex physical AI applications from several months to just a few days.

Palette Neat is an open-source development platform that integrates an execution library and an agent workflow layer.

It is reported that developers can build the entire system using natural language commands and preserve about 90% of existing software assets by reusing existing application code.

The company stated that it reduces the development cycle by directly mapping applications to semiconductors.

The company stated that when this platform is used with Syma AI’s mass-produced Modalix MLSoC System-on-Module (SoM) or PCIe companion card, it can handle workloads in fields such as robotics, automotive, drones, industrial automation, aerospace and defense, smart vision, and healthcare.

Modalix SoM can run multiple large-scale language models (LLM) and vision and sensor models simultaneously with less than 10W of power. Syma AI explained that it is designed as a replacement product that is pin-compatible with the existing NVIDIA SoM form factor, so no separate carrier board redesign is required.

"With Palette Neat and pin-compatible SoMs, we can lower the entry barriers of the existing GPU-centric development ecosystem and enable developers to design systems using natural language, thereby shortening development time to days or hours," said Krishna Rangasayee, founder and CEO of Syma AI.