SiMa.ai
SiMa.ai is a fabless semiconductor company that designs the Modalix Machine Learning System-on-Chip (50 TOPS under 10W) and Palette software suite for Physical AI inference at the edge, serving automotive, industrial, defense, drone, healthcare, and retail OEMs globally through direct sales and regional distributors.
- Company typePrivate
- Founded2018
- HeadquartersSan Jose, United States
- Headcount101–250
- GTM typeB2B
- OfferingHardware or Manufacturing
What SiMa.ai does
SiMa.ai is a fabless semiconductor company founded in November 2018 and headquartered in San Jose, California, that designs the Modalix Machine Learning System-on-Chip (MLSoC) platform purpose-built for Physical AI workloads at the edge. The second-generation Modalix chip, manufactured on TSMC's N6 process with an Arm-based heterogeneous compute architecture and Micron LPDDR5X memory integration, delivers 50 TOPS at under 10 watts of power consumption for computer vision, large language model, vision-language model, and generative AI inference. The company sells its silicon in multiple form factors — Modalix SoM pin-compatible with NVIDIA Jetson Orin NX/Nano, a PCIe HHHL card developed with Advantech, chip-down BGA designs for OEMs, and a DevKit — complemented by the Palette software suite comprising the Palette SDK compiler, Edgematic visual pipeline builder, Model Browser, and the Palette Neat agentic development environment launched in June 2026.
SiMa.ai operates a hybrid business model combining hardware sales (chips, SoMs, PCIe cards, DevKits priced from $449 to $1,499) with software licensing through the Palette SDK. Go-to-market is anchored on direct enterprise sales to automotive OEMs, industrial manufacturers, defense contractors, and drone makers, augmented by an exclusive distribution network — Arrow Electronics for EMEA, Macnica for Japan, and Ortecs for Israel — plus an ODM and system integrator channel that includes Advantech, Enclustra, Kontron, and Lanner. Disclosed design wins span TRUMPF (industrial lasers), Emerson Electric (industrial PCs), Baxter International (medical device manufacturing), ARC Aerospace (defense flight computers), STIGA (robotic lawn mowers), Virya Autonomous Technologies (autonomous mobility), and an unnamed large two-wheeler manufacturer (retail analytics).
The company has raised $355 million in cumulative capital across multiple rounds, most recently an $85 million oversubscribed round led by Maverick Capital in August 2025 and a strategic investment from Micron Technology in April 2026. SiMa.ai maintains offices in eight countries (United States, Germany, India, Israel, Japan, South Korea, Poland, and Taiwan) and operates a wholly-owned India subsidiary. Leadership includes former Xilinx CEO Moshe Gavrielov as Chairman and Intel CEO Lip-Bu Tan as a board member, with Krishna Rangasayee serving as founder and CEO. Revenue is not publicly disclosed and the company remains privately held.
SiMa.ai firmographics
Firmographics- Name
- SiMa.ai
- Legal name
- SiMa Technologies, Inc.
- Website
- https://sima.ai
- Company type
- Private
- Founded year
- 2018
- Operating status
- Operating
- Headcount range
- 101–250 employees
- Short description
- SiMa.ai is a fabless semiconductor company that designs the Modalix Machine Learning System-on-Chip (50 TOPS under 10W) and Palette software suite for Physical AI inference at the edge, serving automotive, industrial, defense, drone, healthcare, and retail OEMs globally through direct sales and regional distributors.
- Ownership category
- akta.pro rank
SiMa.ai industry classification
Industry- Product category
- Edge AI Silicon
- NAICS
- Printed Circuit Assembly (Electronic Assembly) Manufacturing (334418)
- SIC
- Electronic Components, Nec (3679)
- akta.pro primary industry
- Edge AI Hardware & NPUs (SoCs, modules, accelerators) (HDAAAJAC)
- akta.pro secondary industries
- On-Device Inference Runtimes & SDKs (mobile/embedded) (HDAAAJAB), Real-Time Vision at the Edge (on-device CV/video) (HDAAAJAE), AI Application Enablement Platforms (Copilot/Agent Frameworks, SDKs) (HDAEANAJ)
Keywords
Where SiMa.ai is headquartered
LocationHeadquarters
- HQ city
- San Jose
- HQ country
- United States
- HQ region
- North America
Offices8 records
Markets served
SiMa.ai business model
Business model- GTM type
- B2B
- Offering type
- Hardware or Manufacturing
- Cost components
- Technology or R&D, Personnel, Supply Chain, Operations, Marketing or Sales, Infrastructure
Revenue model
- Hardware Sales: Sale of Modalix MLSoC chips, System-on-Modules (SoM), PCIe cards (HHHL), and Development Kits to customers for edge AI deployment.
- Software Licensing: Revenue from Palette SDK software platform including Palette Neat agentic development environment, Model Compiler, and runtime licenses.
- Development Kit Sales: Sale of Modalix DevKit platforms at $1,499 for full development platform with carrier board, SoM, cables, documentation and pre-installed Palette software.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| One time/ perpetual license | Pay-as-you-go | Modalix DevKit - Full development platform |
| One time/ perpetual license | Pay-as-you-go | Modalix SoM - System-on-Module |
| One time/ perpetual license | Pay-as-you-go | Modalix PCIe HHHL Card |
| Other | Multi-year contract | Chip-Down Design |
Go-to-market motion1 record
Distribution channels6 records
Marketing channels6 records
SiMa.ai product offering
Product offeringCore offering
SiMa.ai develops and sells the Modalix Machine Learning System-on-Chip (MLSoC) family, a purpose-built edge AI silicon platform delivering 50 TOPS at under 10W on TSMC N6 with Arm A65 heterogeneous architecture, available as System-on-Modules, PCIe HHHL cards, chip-down BGA packages, and DevKits. The Palette software suite (Palette SDK compiler, Palette Neat agentic development environment, Edgematic visual pipeline builder, Model Browser, and Llima LLM/VLM toolchain) enables ONNX model compilation and deployment of CNNs, transformers, LLMs, VLMs, and Generative AI models at the edge.
Product overview
SiMa.ai offers a unified Physical AI platform combining purpose-built MLSoC hardware with an integrated software suite. The Modalix MLSoC family (second-generation) provides 50 TOPS in under 10W, available as System-on-Module (SoM), PCIe HHHL Card, chip-down BGA, or DevKit. The Palette software platform includes Palette Neat (agentic development environment), Palette SDK (model compiler and toolchain), Edgematic (visual pipeline builder), Model Browser (model selection and download), and Palette Llima (LLM/VLM optimization framework). The platform runs multiple LLMs, VLMs, and CNN models concurrently, enabling multimodal AI at the edge for robotics, automotive, drones, industrial automation, aerospace, defense, smart vision, and healthcare applications.
Differentiator
Problem solved
Functional benefit
Brands
- Modalix: Second-generation Machine Learning System-on-Chip (MLSoC) platform for Physical AI applications, delivering 50 TOPS at under 10W power consumption.
- Palette SDK
- Palette Neat
Products and services
- Modalix MLSoC Family Purpose-built Machine Learning System-on-Chip family delivering 50 TOPS with sub-10W power consumption for Physical AI applications including vision, perception, and autonomous systems. Built on TSMC N6 process with Arm-based heterogeneous architecture supporting CNN, ViT, LLM, and LMM workloads.
- Modalix SoM (System-on-Module) Production-ready system-on-module integrating the Modalix 50 TOPS MLSoC. Pin-compatible with NVIDIA Jetson Orin NX and Nano, enabling drop-in AI upgrades on existing carrier boards without redesign. Available from $449 per unit with volume pricing.
- Modalix PCIe HHHL Card Half-Height Half-Length PCIe add-in card designed for standard commercial and industrial PCs and edge servers. Doubles predecessor performance while operating under 10W. Features PCIe Gen5 x8 interface, passive cooling, and 10 Gigabit Ethernet connectivity.
- Modalix Chip-Down Design Direct PCB integration of Modalix MLSoC for tightest SWaP (Size, Weight, and Power) profile, designed for ODMs and system integrators requiring maximum design freedom. Available in BGA package via contact sales.
- Modalix DevKit Complete hardware development platform with SiMa.ai Palette software pre-installed. Includes carrier board, SoM, cables, and documentation. Available at $1,499 including SoM, PSU, 500GB NVMe M.2, and SiMa.ai Palette SDK.
- Palette SDK Integrated AI application environment and compiler toolchain that compiles any ONNX model directly to SiMa.ai hardware-optimized binary in a single API call. Handles quantization (FP32 to INT8/BF16), layer-wise precision mapping, and memory layout optimization automatically. Supports 200+ ONNX/Hugging Face models including CNN, ViT, LLM, and GenAI architectures.
- Palette Neat Agentic development environment for Physical AI that collapses complex application timelines from months to days or hours. Open-source platform with natural-language interface enabling developers to build systems using plain English commands, with agents autonomously constructing applications and mapping them to silicon hardware. Preserves approximately 90% of legacy software investment.
- Edgematic Visual drag-and-drop development tool for building AI application pipelines. Enables rapid prototyping and deployment of multi-sensor AI applications without writing code, and integrates with AWS for cloud-to-edge deployment.
- Model Browser Interactive web-based tool for exploring and downloading compiled AI models optimized for SiMa.ai Modalix hardware. Provides performance benchmarks and specifications for CNN and LLM/VLM models including Gemma 3, Llama, Mistral, Phi, Qwen, and LLaVA variants.
Quantifiable outcome
- 5x higher YOLOv8n performance than NVIDIA Jetson Orin NX (1,414 FPS vs. competitors)
- +7 more outcomes
Companies that use SiMa.ai
Customer profileNamed customers7 records
Segments6 records
Ideal customer profiles5 records
SiMa.ai technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration3 records
AI capability16 records
Feature9 records
SiMa.ai partnerships and signals
Strategic signalPartnerships
23 partnerships are on record, tiered core and strategic.
- Mistral Solutions (AXISCADES subsidiary)coreStrategic partnership to develop AI-powered autonomous drone reference platform supporting GPS-denied navigation, obstacle detection, sensor fusion, and mission planning. Joint reference design optimized for SWaP constraints enabling sub-10W power operation. Platform targets drone OEMs and system integrators in defense, industrial, and commercial sectors.
- Emerson ElectriccoreCollaboration integrating SiMa.ai's MLSoC technology into Emerson's industrial PCs for factory floors and remote operations. Enables high-performance edge computing without cloud connectivity. Targets industrial automation sector projected to grow from $43.6B (2024) to $153.9B (2030) at 23% CAGR.
- ARC AerospacecorePartnership to develop next-generation advanced flight computers for attritable autonomous systems and low-cost counter-drone missiles. Ruby FC and Ruby CEP series flight computers integrated with MLSoC family for navigation in GPS-denied environments, sensor fusion, and real-time target acquisition.
- FUJISOFTstrategicMOU to accelerate Physical AI deployment in Japan. FUJISOFT developing POC solutions on Modalix architecture and providing hardware/software development support for organizations implementing AI-enabled products.
- Nota AIcoreStrategic partnership combining Nota AI's NetsPresso model compression (90%+ model size reduction) with SiMa.ai's Modalix MLSoC (25-200 TOPS, 10x performance-per-watt). Targets intelligent transportation systems, security/surveillance, and robotics markets.
- AdvantechcorePartnership to develop Modalix PCIe HHHL Card. Advantech contributes industrial IPC platform expertise for seamless integration of AI processing into industrial PCs.
- Images TechnologystrategicStrategic partnership for market entry into Korea. Images provides ultra-precision sensing and actuation technologies including tactile sensing for humanoid robot applications while leveraging SiMa.ai's high-efficiency MLSoC-based AI platform.
- KontronstrategicPartnership to launch KBox A-151 EAI industrial edge AI computer delivering 50+ TOPS performance integrating SiMa.ai's MLSoC Modalix accelerator with 13th Gen Intel Core or Atom processors.
- STIGA S.p.A.strategicPartnership to integrate Physical AI into robotic lawn mowers for domestic and commercial markets. STIGA leverages SiMa's ultra-low-latency, power-efficient MLSoC platform for real-time decision-making in autonomous garden equipment.
- Kaynes SemiconstrategicPartnership to accelerate Physical AI innovation in India and global markets. Focuses on co-developing defense-oriented AI solutions, designing production reference boards for MLSoC platforms, and enabling Make in India electronic systems for industrial automation, robotics, automotive, aerospace, and defense.
- SynopsyscoreFirst integrated joint capability to accelerate automotive AI SoC development. Combined solution enables early architecture exploration and pre-silicon software development for ADAS and IVI applications. Can reduce development costs and shorten vehicle time-to-market by up to 12 months.
- TSMCcoreFoundry partner for Modalix chip manufacturing. Modalix built on TSMC N6 process.
- Cerence AIcorePartnership to power next-gen conversational AI at the edge for automotive applications. CaLLM Edge optimized to run on SiMa.ai MLSoC Modalix platform enabling low-power, low-latency in-car interaction without cloud connectivity.
- L&T Technology Services (LTTS)strategicPartnership to deliver AI-driven solutions across mobility, healthcare, industrial automation, and robotics. LTTS provides engineering expertise while SiMa.ai supplies MLSoC ONE AI hardware and software platforms.
- TRUMPFcorePartnership for AI-supported laser systems for welding, cutting, and marking in manufacturing. TRUMPF leverages SiMa.ai's edge AI platform for intelligent industrial solutions since 2024.
- EnclustrastrategicPartnership to develop pin-compatible System-on-Module using Enclustra's integration expertise and LLiMa Framework for running AI models offline.
- Macnica (Japan)coreStrategic distribution agreement to accelerate Physical AI adoption in Japan market.
- CiscostrategicPartnership to bring low-latency, energy-efficient AI to Industry 4.0 by combining SiMa.ai's Modalix AI platform with Cisco's IE3500 industrial Ethernet switches.
- Wind RiverstrategicCollaboration to enable seamless AI/ML for intelligent edge applications. Integration combines Wind River's software stack with SiMa.ai's MLSoC for mission-critical aerospace, defense, robotics, and industrial automation deployments.
- Ortecs (Israel)strategicRepresentative and distribution agreement for MLSoC products in Israel.
- Arrow ElectronicscoreExclusive distribution agreement in EMEA region for MLSoC, Palette Software and Development Kits. Arrow serves as region's exclusive distributor.
- ArmcoreTechnology partner providing Arm-based architecture (8-core 1.5GHz Arm A65 microprocessor) integrated into Modalix MLSoC. SiMa.ai participates in Arm Flexible Access licensing program for edge AI platform.
- Virya Autonomous TechnologiesstrategicPartnership to bring powerful edge-AI capabilities for real-time, low-power decision-making enabling smarter, faster, and scalable solutions for India's autonomous mobility space.
Scale indicators4 records
Recent moves7 records
Expansion highlights7 records
SiMa.ai competitors and assessment
Company assessmentDirect peers
- NVIDIA: NVIDIA's Jetson Orin NX/Nano is the explicit pin-compatibility target for SiMa.ai's Modalix SoM, and both companies compete head-on for low-power on-device AI inference in robotics, industrial, automotive, and edge AI workloads.
- Hailo: Hailo builds edge AI accelerators (Hailo-8, Hailo-15) for the same industrial, automotive, and vision workloads as Modalix, with similar performance-per-watt positioning and overlap in camera and industrial PC partners.
- Axelera AI: Axelera AI develops edge AI accelerators (Metis) targeting computer vision and GenAI workloads at the edge — a near-direct competitor on inference TOPS/W and Model Browser-style developer tooling.
- Graphcore: Graphcore develops IPUs for low-latency AI inference, with edge-focused positioning for industrial and enterprise workloads that overlap Modalix's target use cases and developer-tooling strategies.
Broad incumbents
- Qualcomm: Qualcomm's IoT and Snapdragon Ride/Automotive platforms offer on-device AI for industrial, automotive, and IoT, competing with Modalix across smart cameras, robotics, and ADAS adjacencies as part of a much broader silicon portfolio.
- Intel (Movidius / OpenVINO): Intel's Movidius VPU line and OpenVINO ecosystem serve edge inference across industrial, retail, and surveillance — overlapping Modalix's industrial PC and vision workloads as part of a much larger compute portfolio.
- AMD (Xilinx): AMD/Xilinx's Kria SOM and Versal adaptive SoCs target embedded vision and edge AI in industrial and robotics — the same SiMa.ai verticals — as part of AMD's broader FPGA and adaptive compute portfolio.
Emerging players
- Cerebras Systems: Cerebras primarily targets data-center scale AI but is increasingly competing for on-device and Physical AI workloads; both companies are profiled together in edge AI infrastructure rankings.
- Tenstorrent: Tenstorrent builds RISC-V-based AI accelerators targeting edge and data center; they overlap with Modalix on custom-silicon development ecosystems and developer-tools positioning for low-power AI.
Others
- Synopsys: Synopsys is both a SiMa.ai partner (joint automotive AI SoC blueprint) and an EDA ecosystem enabler whose IP is used in competitors' edge AI silicon; tangential rather than direct overlap but ecosystem-level comparability.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat6 records
Key risks6 records
Key highlights7 records
Customer concentration
SiMa.ai social profiles
Digital presenceSiMa.ai compliance and trust
Trust signalCompliance2 records
SiMa.ai financial estimates
Financial estimateRevenue estimate
Valuation estimate
SiMa.ai leadership team
Management profileNumber of profiles
Profiles14 records
SiMa.ai subsidiaries and ownership
Company hierarchySubsidiaries1 record
SiMa.ai funding detail
Funding detailFunding overview
Funding rounds11 records
Investors18 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
SiMa.ai M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about SiMa.ai
What does SiMa.ai do?
SiMa.ai develops and sells the Modalix Machine Learning System-on-Chip (MLSoC) family, a purpose-built edge AI silicon platform delivering 50 TOPS at under 10W on TSMC N6 with Arm A65 heterogeneous architecture, available as System-on-Modules, PCIe HHHL cards, chip-down BGA packages, and DevKits. The Palette software suite (Palette SDK compiler, Palette Neat agentic development environment, Edgematic visual pipeline builder, Model Browser, and Llima LLM/VLM toolchain) enables ONNX model compilation and deployment of CNNs, transformers, LLMs, VLMs, and Generative AI models at the edge.
Is SiMa.ai a public or private company?
SiMa.ai is a private company. It is classified as venture growth investor backed and is currently operating.
When was SiMa.ai founded?
SiMa.ai was founded in 2018. It employs 101 to 250 people.
Where is SiMa.ai based?
SiMa.ai is headquartered in San Jose, United States, in the North America region.
How does SiMa.ai make money?
Three revenue lines are on record. Hardware Sales are the primary driver. The others are software Licensing and development Kit Sales.
Who are SiMa.ai's main competitors?
Direct peers on record are NVIDIA, Hailo, Axelera AI and Graphcore. Broad incumbents are Qualcomm, Intel (Movidius / OpenVINO) and AMD (Xilinx). Emerging players are Cerebras Systems and Tenstorrent. Synopsys is listed as an others.
Does SiMa.ai have an API?
Yes. SiMa.ai provides a Palette SDK for model compilation and deployment. The ModelSDK enables developers to load, quantize, and compile ONNX models to SiMa.ai hardware-optimized binaries. Developers can build complete pipelines using Python, C++, or Gstreamer APIs. Edgematic provides a visual pipeline builder for drag-and-drop application development. The Palette Neat agentic environment enables natural-language-based development workflows. Documentation and developer resources are available through the Developer Center at developer.sima.ai. Developer documentation is at developer.sima.ai/?moved=true.
What industry is SiMa.ai in?
SiMa.ai's product category is Edge AI Silicon. Its primary akta.pro industry code is HDAAAJAC, Edge AI Hardware & NPUs (SoCs, modules, accelerators), with a secondary code of HDAAAJAB, On-Device Inference Runtimes & SDKs (mobile/embedded). Its NAICS code is 334418 and its SIC code is 3679.