cd /news/ai-chips/sima-ais-palette-3-0-runs-many-ai-mo… · home › topics › ai-chips › article
[ARTICLE · art-171065] src=siliconangle.com ↗ pub= topic=ai-chips verified=true sentiment=↑ positive

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

SiMa Technologies Inc. released Palette 3.0, a software suite that runs multiple AI models concurrently on a single Modalix MLSoC chip and raises model throughput by up to 3.5 times, the company said. The throughput gain comes from speculative decoding using the EAGLE3, DFlash and Gemma 4 MTP algorithms, and the release adds support for large mixture-of-experts models including OpenAI's GPT-OSS-20B on the Modalix MLSoC 16GB system-on-module. SiMa.ai, founded in 2018 by CEO Krishna Rangasayee and backed by a $150 million Series C at a $1.45 billion valuation in September, targets drones, robots and other autonomous machines that must run AI models on-device rather than in a data center.

by read3 min views1 publishedNov 17, 2026
SiMa.ai’s Palette 3.0 runs many AI models on one chip and lifts throughput up to 3.5 times
Image: Siliconangle (auto-discovered)

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 as much as 3.5 times.

Developers building drones, robots and other autonomous machines, which need to run AI models on the device rather than in a data center, are a primary focus of the release. SiMa.ai argues that while industry incumbents rely on adapting legacy data center technology, it builds its software and silicon exclusively for physical AI.

Leading the new capabilities in Palette 3.0 is improved concurrent model execution on SiMa.ai’s Modalix MLSoC. Maintaining what the company describes as a low-power profile, a single chip can handle complex reasoning and multimodel workloads at the same time. The platform can also run separate Docker containers on the chip. Supported workloads include large language models and vision-language models.

Perception, depth and reasoning models used by drones and robots are also said to run faster on the new version. Working with ecosystem partners, SiMa.ai has added unified sensor fusion, dense neural depth fields and closed-loop digital twin simulation to the platform. The goal of the additions, SiMa.ai says, is to remove the integration and validation bottlenecks that slow down autonomous development.

Performance figures for those models are published in a new Model Registry, which SiMa.ai says measures every result within real-world power constraints. The company is offering competitive benchmarking details on request.

Also new in the release is a set of open-source capabilities for generative AI application development. Behind the 3.5-times throughput figure is speculative decoding, using the EAGLE3, DFlash and Gemma 4 MTP algorithms. Support has also been added for large mixture-of-experts models such as OpenAI Group PBC’s GPT-OSS-20B, which can now run on the Modalix MLSoC 16GB system-on-module.

Developers are free to adapt the open-source code, which is available on GitHub. Documentation for Palette 3.0 is provided through SiMa.ai’s Developer Center.

Founded in 2018 by Chief Executive Krishna Rangasayee, SiMa.ai offers a platform that combines purpose-built silicon with software for AI on physical devices. Notable customers and partners include Robert Bosch GmbH, Emerson Electric Co. and Synopsys Inc.

SiMa.ai raised $150 million in a Series C round at a $1.45 billion valuation in September. Fidelity Management & Research Co. and Amplify Partners co-led the round. The company has raised $500 million to date.

Image: SiMa.ai

Support our mission to keep content open and free by engaging with theCUBE community. Join theCUBE’s Alumni Trust Network, where technology leaders connect, share intelligence and create opportunities.

  • 15M+ viewers of theCUBE videos , powering conversations across AI, cloud, cybersecurity and more
  • 11.4k+ theCUBE alumni — Connect with more than 11,400 tech and business leaders shaping the future through a unique trusted-based network

Are you an AWS customer?  Support SiliconANGLE financially by buying your AWS services from our Marketplace portal page and links: https://siliconangle.com/aws-marketplace/

About SiliconANGLE Media

SiliconANGLE,

theCUBE Network,

theCUBE Research,

CUBE365,

theCUBE AIand theCUBE SuperStudios — with flagship locations in Silicon Valley and the New York Stock Exchange — SiliconANGLE Media operates at the intersection of media, technology and AI.

Founded by tech visionaries John Furrier and Dave Vellante, SiliconANGLE Media has built a dynamic ecosystem of industry-leading digital media brands that reach 15+ million elite tech professionals. Our new proprietary theCUBE AI Video Cloud is breaking ground in audience interaction, leveraging theCUBEai.com neural network to help technology companies make data-driven decisions and stay at the forefront of industry conversations.

── more in #ai-chips 4 stories · sorted by recency
── more on @sima technologies inc. 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/sima-ais-palette-3-0…] indexed:0 read:3min 2026-11-17 · —