SiMa.ai Raises $150 Million to Scale Chips for Humanoids, Physical AI
The semiconductor company plans to use its latest financing to develop higher-performance edge-AI hardware and software for physical systems, including humanoid robots.

SAN JOSE, Calif. — SiMa.ai raised $150 million in Series C financing as the semiconductor company develops a new generation of computing hardware for physical AI hardware, including humanoid robots.
The Series C round, which values SiMa.ai at $1.45 billion, was co‑led by Fidelity Management & Research Co. and Amplify, increases its total funding to $500 million, the company announced this week. Existing investors including Dell Technologies Capital, Maverick Capital, Point72 and StepStone Group participated. AllianceBernstein, Baron Capital, J.P. Morgan and the State of Michigan joined as new investors.
The money is not just for robotics. SiMa.ai develops chips and software for physical-AI applications spanning robots, vehicles, drones and other edge systems. But humanoid robots are a specific focus of the company's next hardware-and-software push, and the new funding will help finance its third-generation platform.
The company is aiming to solve a key computing bottleneck for humanoid robots: running demanding perception, reasoning and control models locally while keeping power and latency low.
Its existing Modalix machine-learning system-on-chip is designed to run AI inference directly on robots. SiMa.ai says the platform can support foundation model operations for humanoid robots at the edge, while its robotics systems perform AI inference under a power budget of less than 10 watts. These are company performance claims without independent benchmarking.
The company has already used Modalix to run Nvidia's GR00T robot foundation-model stack. A SiMa.ai technical brief says the implementation merges perception and reasoning with robot action into one system-on-chip to eliminate cloud processing during operation.
The Series C will fund a more powerful generation of hardware. SiMa.ai is targeting 1,000 dense tera operations per second, or TOPS, for that portfolio, with availability planned for the first half of 2028. This is still a development benchmark, not a deliverable metric.
SiMa.ai has not disclosed how much of the $150 million will be spent specifically on humanoid robotics, nor has it announced a humanoid manufacturer as a customer for the planned third-generation hardware.
More articles on this topic
71 articles
ReportAmazon Unit Connects Nvidia Simulation With Robot Training and Deployment
ReportUBTECH, FAW-Volkswagen Expand Humanoid Robot Work Into Logistics
ReportMecka Raises $60 Million as Demand for Robot Training Data Grows
ReportMinerva Humanoids Raises $10 Million for Robots Built for Dangerous Jobs
ReportAsmech Starts Selling $6,999 Elephant Robotics Cobot in Japan
ReportKUKA Taps TD SYNNEX to Expand Mobile Robot Sales in Europe
ReportBoston Dynamics Puts Former Amazon Executive at Helm for Physical AI Push
ReportHumansX launches Wuji dexterous hand and data glove in Japan
ReportDoosan Gets $51M in State Support for Physical AI
ReportTeradyne Invests in Bright Machines to Link Robots and Factory Data
ReportPanasonic Targets Humanoid Robot Production by 2029
ReportPersona AI Adds Daiwa House Fund as Industrial Humanoid InvestorBUSINESS NEWS WEEKLY LETTER
The Weekly Letter for Robotics Professionals, Summarizing the Most Important Industry Moves, Launches, Deals and Signals.