SiMa.ai

Edge AI

SiMa.ai is a San Jose–based semiconductor and software company founded in 2018 that develops machine learning system-on-chips (MLSoCs) and software platforms for running AI inference at the edge—in robots, industrial equipment, automotive systems, and smart devices—without cloud dependency.

Reviewed by 7wData
Share Email

On this page

Profile

Builds power-efficient AI chips and software platforms that let devices run machine learning models at the edge without sending data to the cloud.

SiMa.ai is a San Jose–based semiconductor and software company founded in 2018 that develops machine learning system-on-chips (MLSoCs) and software platforms for running AI inference at the edge—in robots, industrial equipment, automotive systems, and smart devices—without cloud dependency. The company was founded by Krishna Rangasayee, formerly COO of Groq and a 18-year veteran of Xilinx where he held senior executive roles. Moshe Gavrielov, the former president and CEO of Xilinx, chairs SiMa.ai's board.

SiMa.ai has raised approximately $345 million across eight funding rounds since 2020, backed by investors including Fidelity Management & Research, Maverick Capital, Dell Technologies Capital, MSD Partners, Point72, and Amplify Partners. The most recent public funding event was an $85 million Series C-II round in August 2025. In April 2026, the company announced a strategic investment from memory chipmaker Micron Technology, strengthening collaboration around integrated compute-memory architectures for edge AI, though the investment amount was not disclosed.

The company's flagship product is the Modalix MLSoC, a second-generation machine learning processor that delivers up to 50 TOPS of performance while consuming under 10 watts. SiMa.ai launched Modalix into production in August 2025 with pricing starting at $349 for an 8GB system-on-module and $599 for 32GB. The accompanying software framework, LLiMa, enables deployment of large language models, vision transformers, and generative AI models directly on the chip. The company has announced approximately 220 to 235 employees as of mid-2025.

Recent strategic moves include expanded collaboration with Synopsys announced in July 2025 to develop automotive AI IP and reference SoC designs, with production availability targeted for end of 2026. In May 2026, SiMa.ai partnered with Emerson to integrate its MLSoC into industrial edge PCs for factory automation, predictive maintenance, and real-time quality inspection. SiMa.ai positioned itself not as a competitor to NVIDIA but as a precision tool for low-power edge inference, claiming to outperform NVIDIA's Orin in certain MLPerf benchmarks. The company was recognized with the Edge AI + Vision Alliance 2026 Product of the Year award for Modalix SoM.

Track SiMa.ai and 240+ vendors.

48k+ subscribers read the daily AI & data note. One email, both newsletters. Unsubscribe anytime.

Who buys this

  • Industrial automation and smart manufacturing (OEMs, integrators running predictive maintenance and quality control)
  • Robotics and autonomous systems developers (mobile robots, AMRs, industrial equipment)
  • Automotive OEMs and Tier-1 suppliers (ADAS, autonomous driving stacks, in-vehicle intelligence)
  • Edge computing appliance vendors and system integrators (Emerson, Lanner, Supermicro, enclosure builders)
  • Government, defense, and aerospace contractors requiring on-premise inference (no cloud dependency)

Publicly disclosed clients

  • Emerson (industrial edge PCs)
  • Synopsys (automotive IP and design tools partnership)
  • Lanner (edge AI appliances)
  • Baxter International (mentioned in company materials)
  • Trumpf Laser Technology (mentioned in company materials)

Strengths and what to watch

Strengths

  • Proven power efficiency advantage in edge inference: MLPerf Inference 3.0 benchmarks show Modalix outperforming NVIDIA Orin in closed-edge ResNet50, and 85% better efficiency in multi-stream power benchmarks
  • Veteran founding team with silicon industry credibility: founder Krishna Rangasayee spent 18 years at Xilinx as SVP, bringing enterprise sales and go-to-market experience; board chair Moshe Gavrielov led Xilinx to $2.5B+ revenue
  • Strong institutional backing with strategic partners: $345M+ in funding from tier-one VCs (Fidelity, Amplify, MSD) plus strategic investments from Micron; ecosystem partnerships with Synopsys, Arm, TSMC, and Emerson validate the platform

Watch for

  • Capital intensity and path to profitability unclear: $345M raised suggests substantial burn; semiconductor fab relationships (TSMC) and software licensing model need monitoring to assess when unit economics turn positive
  • Automotive timeline execution risk: Synopsys collaboration promised ML IP availability mid-2026 and production mid-2027; delays would signal execution challenges in a key target vertical
  • NVIDIA's dominance in adjacent markets: while SiMa targets sub-25W edge inference, NVIDIA's aggressive edge positioning (Orin Nano, NX pricing) and software ecosystem (CUDA, TensorRT) remain formidable competitive moats

Recent moves

Key Information

Industry
Edge AI
Founded
2018
Headquarters
San Jose, USA

Frequently Asked Questions

What does SiMa.ai do?

SiMa.ai builds machine learning system-on-chips (MLSoCs) and software platforms enabling AI inference at the edge without cloud dependency. Founded in 2018 by former Xilinx and Groq executives, the San Jose company targets robotics, industrial equipment, automotive systems, and smart devices with power-efficient hardware.

What is Modalix and how much does it cost?

Modalix is SiMa.ai's second-generation MLSoC delivering 50 TOPS of performance consuming under 10 watts. Pricing starts at $349 for the 8GB system-on-module and $599 for 32GB, launched in production August 2025. The LLiMa software framework enables deployment of large language models and generative AI directly on the chip.

How much funding has SiMa.ai raised?

SiMa.ai has raised approximately $345 million across eight funding rounds since 2020 from investors including Fidelity, Maverick Capital, Dell Technologies, and Point72. The most recent was an $85 million Series C-II round in August 2025, plus a strategic investment from Micron Technology in April 2026 for integrated compute-memory architectures.

How does SiMa.ai compare to NVIDIA?

SiMa.ai positions itself not as a competitor to NVIDIA but as a precision tool for low-power edge inference. Modalix outperforms NVIDIA's Orin in MLPerf Inference 3.0 closed-edge ResNet50 benchmarks and delivers 85% better efficiency in multi-stream power tests. Both target different market segments within edge AI.

What industries use SiMa.ai's chips?

SiMa.ai targets industrial automation, robotics, automotive OEMs, government and defense contractors. Key applications include predictive maintenance, quality control, autonomous systems, ADAS, real-time anomaly detection, and on-premise inference requiring no cloud dependency. Notable partners include Emerson, Synopsys, Lanner, and Baxter International for industrial edge computing solutions.

Is SiMa.ai developing automotive AI chips?

SiMa.ai expanded its Synopsys collaboration to develop automotive AI IP and SoC designs targeting mid-2026 early access and end-2026 production. This accelerates ADAS and autonomous driving integration. In May 2026, SiMa.ai partnered with Emerson to integrate Modalix into industrial edge PCs for autonomous safety, predictive maintenance, and real-time quality inspection.

How SiMa.ai compares

Direct head-to-head against 3 competitors. Picked by 7wData.

This company

SiMa.ai

Positioning
Builds power-efficient AI chips and software platforms that let devices run machine learning models at the edge without sending data to the cloud.
Customer segments
Industrial automation and smart manufacturing (OEMs, integrators running predictive maintenance and quality control)
Strengths
Proven power efficiency advantage in edge inference: MLPerf Inference 3.0 benchmarks show Modalix outperforming NVIDIA Orin in closed-edge ResNet50, and 85% better efficiency in multi-stream power benchmarks
Watch for
Capital intensity and path to profitability unclear: $345M raised suggests substantial burn; semiconductor fab relationships (TSMC) and software licensing model need monitoring to assess when unit economics turn positive
Recent moves
Emerson and SiMa.ai partner to deploy Physical AI in industrial edge PCs with autonomous safety, predictive maintenance, and real-time anomaly detection

Hailo

Positioning
Israeli edge AI chip maker targeting automotive cockpit, driver monitoring, and security camera inference at sub-25W power.
Customer segments
Automotive OEMs, security camera makers, consumer device builders, and industrial integrators; 300 plus customers across 40 countries.
Strengths
Raspberry Pi AI Kit partnership (June 2024) placed Hailo-8 in a mass developer ecosystem no other pure-play edge inference startup accessed.
Watch for
Valuation dropped from $1.2B to under $500M; January 2026 layoffs cut 10 percent of workforce; emergency $9M loan signals acute cash distress.
Recent moves
April 2026: non-binding SPAC merger signed targeting public listing at under $500M valuation after major investor wrote down stake by $77M.

Ambarella

Positioning
Public semiconductor company with edge AI SoCs in volume production across automotive ADAS, robotics, and enterprise security cameras.
Customer segments
Automotive OEMs and Tier-1 suppliers, drone manufacturers, industrial robotics builders, enterprise security vendors; 370 plus production AI projects.
Strengths
45 million plus cumulative edge AI SoCs shipped; CV3-AD on 5nm entered mass automotive production in 2025 with design wins across China and Europe.
Watch for
Automotive and enterprise security targeted to exceed 85 percent of revenue; ADAS spending slowdowns or China customer concentration create earnings volatility.
Recent moves
January 2026: launched CV7 8K edge AI SoC at CES targeting robotics, multi-stream automotive, and enterprise security; FY2026 revenue reached $390.7M, up 37 percent.

Axelera AI

Positioning
Dutch edge AI chip startup targeting European enterprise and defense buyers with high-TOPS inference silicon as a non-US supply alternative.
Customer segments
Defense contractors, retail and agriculture enterprises, industrial robotics builders; 350 plus enterprise customers skewing European.
Strengths
Europa AIPU rated at 629 TOPS at 45W with over 500 customers signed before H1 2026 general shipment start.
Watch for
Europa chip sampling started only in Q2 2026; field-proven production deployments remain limited versus SiMa and Ambarella already shipping at volume.
Recent moves
February 2026: raised $250M plus led by Innovation Industries, BlackRock, and Samsung Catalyst Fund; largest EU semiconductor private funding round on record.

Sources

  1. techcrunch.com — April 2024 $70M funding round led by Maverick Capital; total funding trajectory to $270M; intended use of funds for second-gen MLSoC and go-to-market expansion
  2. sima.ai — Modalix SoM announcement date (March 10, 2025); 50 TOPS performance; features and use cases; partnership with Enclustra; Early Access Program
  3. www.businesswire.com — October 2022 Series B1 funding extended to $67M with MSD Partners; cumulative $187M raised; list of existing investors (Fidelity, Amplify, Dell, Wing, +ND, Lip-Bu Tan, Alter)
  4. www.businesswire.com — May 2020 Series A $30M round led by Dell Technologies Capital; participation from Amplify Partners, Wing Venture Capital, +ND Capital; founder and team background
  5. siliconangle.com — August 2025 Modalix production launch; pricing ($349 for 8GB, $599 for 32GB SoM, $1,499 DevKit); LLiMa framework; CEO quote on demand; availability details
  6. sima.ai — April 8, 2026 Micron strategic investment announcement; focus on LPDDR5X memory integration; CEO Krishna Rangasayee quote on performance/power efficiency
  7. sima.ai — July 30, 2025 Synopsys collaboration expansion; automotive IP and reference SoC designs; Platform Architect integration; mid-2026 early access timeline, end-2026 production, mid-2027 chiplet target
  8. www.prnewswire.com — May 26, 2026 Emerson partnership announcement; industrial PC integration; autonomous safety, predictive maintenance, quality inspection use cases; quote from CEO on closed-loop autonomy; market context on industrial AI growth