SiMa.ai, a startup building chips and software that let robots, drones, and cameras run AI workloads directly on the device, has closed a $150 million Series C at a $1.45 billion valuation. Fidelity Management & Research Company and Amplify co-led the round, with Alter Venture Partners, Dell Technologies Capital, and StepStone Group participating.

The deal marks a sharp step up from the company's last raise. SiMa.ai was valued at $960 million after its $85 million Series B in July 2025, according to PitchBook. The new round pushes total capital raised past $500 million — a 51% valuation jump in under a year, achieved while edge AI silicon remains one of the most crowded segments in semiconductors.

Krishna Rangasayee founded SiMa.ai in 2018. He previously served as COO of Groq, the inference-chip specialist that has itself become one of the most heavily funded AI silicon startups. That lineage matters: Rangasayee's pitch is not a general-purpose accelerator but a power- and cost-optimized part aimed squarely at embedded devices.

The company's core argument is energy efficiency. Its chips run AI inference locally, eliminating the round trip to the cloud that conventional deployments require. For battery-powered and thermally constrained systems — drones, mobile robots, smart cameras — that local execution cuts latency and power draw at the same time. SiMa.ai is betting that this combination, plus pricing it describes as more affordable than Nvidia's GPUs, will win sockets in the emerging market for physical AI devices, including humanoid robots.

The competitive framing against Nvidia is deliberate and telling. Nvidia dominates AI compute in the data center, where power budgets are measured in hundreds of watts and cost per module is a secondary concern. Edge devices invert those constraints: a humanoid robot or an inspection drone cannot carry an H-class GPU, and even Jetson-class modules can strain the bill of materials at scale. That gap is what SiMa.ai and a field of edge-AI rivals — Hailo, Ambarella, Qualcomm's robotics line, and Intel's Movidius descendants — are all chasing.

The investor list signals where the company sees its nearest-term demand. Dell Technologies Capital's participation ties SiMa.ai closer to a major systems vendor at a moment when PC and server makers are racing to add on-device AI. Fidelity's involvement, meanwhile, gives the startup a marquee institutional backer of the kind that typically precedes either a late-stage round or an IPO path.

The total funding picture — more than $500 million raised across four years of Series-stage rounds — puts SiMa.ai in the second tier of AI silicon startups by capital raised, well below the multibillion-dollar war chests of Groq or Cerebras, but ahead of most embedded-AI specialists. Its $1.45 billion valuation clears the unicorn threshold without the revenue multiples that data-center AI chipmakers have commanded.

The open question is design wins. The company has not disclosed specific customer volumes, named design partners, or shipment figures alongside the raise. Whether the humanoid robot market materializes at commercial scale — and how quickly device makers consolidate around one or two edge-inference platforms — will determine whether this valuation holds or compresses in the next round. SiMa.ai now has the capital to wait for that market; its investors are betting roughly half a billion dollars that it arrives on schedule.