SiMa.ai builds chips and software that let developers run AI directly inside robots, drones and vehicles, where machines must turn sensor data into actions. The San Jose startup announced a $150 million Series C at a $1.45 billion valuation, funding a push to broaden that embedded-computing platform.
Instead of sending every task to a cloud data center, SiMa.ai’s platform runs models on the device itself. A drone or car can combine data from cameras and other sensors with an AI model, then act through the machine’s controls. SiMa.ai is targeting drones, humanoid robots, advanced driver-assistance systems and AI-powered cockpits.
The stack pairs the Modalix system-on-chip with Palette Neat, a development environment for building and deploying physical-AI software. SiMa.ai says the combination can reduce deployment work from months to days or hours. Part of the new capital will expand Palette Neat and fund another hardware generation.
This is a power-and-software bet as much as a chip bet. SiMa.ai markets its platform as a cheaper, lower-power alternative to Nvidia GPUs and their CUDA software ecosystem, according to TNW. Its next hardware is planned for the first half of 2028, with a company target of 1,000 dense trillion operations per second.
Commercial traction is less clear. SiMa.ai said revenue growth quadrupled from 2024 to 2025 but disclosed no revenue figures. It lists Bosch, Emerson, Micron, Synopsys and TRUMPF among customers and partners, without distinguishing which are buyers, technical partners or both.
Fidelity Management & Research and Amplify co-led the round. TNW said it brings total funding to $500 million and cited earlier TechCrunch reporting, based on PitchBook data, that SiMa.ai raised an $85 million Series B at a $960 million valuation in July 2025. On that basis, the latest valuation is about 51% higher.
Published versions remain available when an article is updated or corrected.
- Revision 1 · Initial publication
Initial publication