Urgent.News

What's breaking now, across thousands of outlets.

AI

SiMa.ai’s Palette 3.0 runs many AI models on one chip and lifts throughput up to 3.5 times

Physical artificial intelligence chip startup SiMa Technologies Inc. today released Palette 3.0, a new version of its Palette software suite intended to allow many AI models to run concurrently on one chip. According to the company, the release, which builds on the Palette Neat agentic development environment launched in June, can raise model throughput by […] The post SiMa.ai’s Palette 3.0 runs…

SiMa.ai’s Palette 3.0 runs many AI models on one chip and lifts throughput up to 3.5 times

Physical AI chip startup SiMa Technologies Inc. unveiled Palette 3.0, an enhanced version of its Palette software suite, designed to enable multiple AI models to run simultaneously on a single chip. The upgrade, which leverages the company's Palette Neat agentic development environment, promises to increase model throughput by up to three and a half times. Developers working on drones, robots, and other autonomous machines that require on-device AI models are the primary beneficiaries of this release.

SiMa.ai emphasizes that while competitors rely on adapting legacy data center technology, their software and hardware are specifically tailored for physical AI. The new Palette 3.0 features improved concurrent model execution on their Modalix MLSoC, ensuring complex reasoning and multitasking capabilities while maintaining a low-power profile. The platform can run separate Docker containers on the chip and support workloads such as large language models and vision-language models.

To address integration and validation bottlenecks that can slow down autonomous development, SiMa.ai has incorporated unified sensor fusion, dense neural depth fields, and closed-loop digital twin simulation into the platform. Performance metrics for these models can be found in the new Model Registry, which measures all results within real-world power constraints. Benchmarks and details on this improvement can be obtained upon request.

Another addition to the release is a suite of open-source capabilities for generative AI application development. These enhancements employ the EAGLE3, DFlash, and Gemma 4 MTP algorithms, and SiMa.ai invites developers to adapt the open-source code available on GitHub. Comprehensive documentation for Palette 3.0 is accessible through SiMa.ai's Developer Center.

SiMa.ai, founded in 2018 by CEO Krishna Rangasayee, provides a platform that merges specialized hardware with software for AI on physical devices. Notable customers and partners include Robert Bosch GmbH, Emerson Electric Co., and Synopsys Inc. The company has raised $500 million in total funding, with $150 million secured in a September Series C round led by Fidelity Management & Research Co. and Amplify Partners.

Written by urgent.news from SiliconANGLE's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at siliconangle.com →

More in AI

More from Tuesday 17 November →