SiMa.ai raises $150 million at a $1.45 billion valuation as the race for physical AI chips moves to the edge of devices

SiMa.ai closed a Series C financing round of $150 million, raising its valuation to $1.45 billion and entering the billionaire club. The round was led by Fidelity Management & Research and Amplify, with participation from Alter Venture Partners, Dell Technologies Capital and StepStone Group. With this round, the company’s total funding exceeds $500 million, after its market value reached $960 million following a previous $85 million financing in July 2025, according to PitchBook data.

The company was founded in 2018 by Krishna Rangasayi, who previously served as chief operating officer at the chip maker Groq. It focuses on developing custom chips and software that enable AI models to run directly on physical devices such as robots, drones, smart cameras and autonomous equipment, without sending data back and forth to cloud servers. The company bets on delivering high energy efficiency and low-latency response, together with prices more affordable than Nvidia graphics processors, to capture a share of the expanding physical AI and humanoid-robot market.

Shifting from cloud centralisation to the edge represents a structural change in AI economics.Relying exclusively on data processing inside cloud data centres imposes growing operational and financial constraints, starting with the cost of moving massive data and extending to unacceptable latency in critical motion-sensitive applications; when machines interact with the real world through computer vision and instantaneous sensing, local inference on the chip becomes essential to ensure safe movement and uninterrupted operation, independent of network fluctuations.

This shift reshapes the choices of engineering teams and tech leaders in the Gulf, Egypt and the Levant on several levels: first, infrastructure costs in smart-city projects, industrial-site monitoring and ports, where running AI on cameras and surveillance devices eliminates the need to pay for streaming live video to the cloud and reduces reliance on expensive graphics processors in distant data centres; second, data localisation requirements, as on-chip inference keeps sensitive data inside the facility without traversing public networks; it also opens the door for robot and embedded-system engineers in Cairo, Amman and Riyadh to build drones and field robots with lower hardware costs and reduced power consumption, without hitting the scarcity of cloud processors or their high licensing fees.

In our coverage of infrastructure evolution, we see capital flowing into specialised edge-inference chips putting an end to the unified-solution hypothesis based on stacking cloud GPUs. Builders of physical AI systems in the region now need to reassess the viability of central hosting and weigh cloud-server costs against the efficiency of local chips that can manage operational decisions on site with minimal latency and optimal energy use.