SiMa.ai Raises $150M for Physical AI Scale

SiMa.ai has raised a $150M Series C at a company-stated $1.45B valuation to expand a Physical AI platform built around Palette software and Modalix machine-learning silicon. Fidelity Management & Research Company and Amplify co-led the September 28, 2026 financing, with a mix of returning and new investors supporting the San Jose company.

The round is aimed at a practical constraint inside robotics, drones, vehicles, industrial systems, and smart machines: the model may be intelligent, but the product still has to make decisions locally under tight limits on power, latency, heat, safety, and integration time. SiMa.ai is betting that software and silicon designed together can make that handoff easier.

SiMa.ai says the financing brings total capital raised to $500M. That total does not fully reconcile with the company's prior disclosure of $355M raised before the new $150M round, so the cumulative figure should be read as the company's current reported total rather than an independently audited sum.

What SiMa.ai Announced

The Series C announcement names Fidelity Management & Research Company and Amplify as co-leads. Alter Venture Partners, Dell Technologies Capital, Maverick Capital, +ND Capital, Point72, and StepStone Group participated, while AllianceBernstein, Baron Capital, J.P. Morgan, and the State of Michigan joined as new investors.

The financing values SiMa.ai at $1.45B, although the release does not say whether that figure is pre-money or post-money. It follows an $85M round announced in August 2025 and an additional $70M announced in April 2024. Founder and CEO Krishna Rangasayee started SiMa.ai in 2018 after senior semiconductor roles that included Xilinx, Groq, and Altera.

SiMa.ai is headquartered in San Jose and says it now operates across eight countries. The company's current markets include robotics, automotive, drones, industrial automation, aerospace and defense, smart vision, and healthcare, all of which put computing into places where a cloud round trip may be too slow, too costly, or unavailable.

The Product Bet Behind the Round

SiMa.ai's platform combines its Palette software stack with the Modalix MLSoC family. Palette Neat is an agentic development environment designed to let engineering teams build Physical AI applications through natural-language and conventional programming workflows. The Modalix platform carries those applications onto purpose-built edge hardware.

That combination matters because a new accelerator is rarely a drop-in business decision. An engineering team has models, sensors, application code, safety requirements, operating systems, carrier boards, procurement plans, and field-support obligations already attached to a product. Better compute can lose the deal if adopting it forces the customer to rebuild everything around the chip.

SiMa.ai says Palette Neat can reduce development from months to days or hours and preserve much of a customer's existing application investment. The company also says its production Modalix system-on-module can run language, vision, and sensor models under 10W. Those are company-reported product claims, but they describe the right commercial target: make the migration cost small enough that power and performance gains can survive the integration meeting.

Where the $150M Is Going

SiMa.ai plans to use the Series C to scale Palette Neat and fund a new generation of machine-learning IP, chiplets, and systems-on-chip. The roadmap targets 1,000 dense TOPS of compute and first-half 2028 availability. Medium- and high-end drones, humanoid robotics, advanced driver-assistance systems, and AI-powered vehicle cockpits are the named priority markets.

Those targets are forward-looking rather than delivered specifications. The capital gives SiMa.ai more room to build the hardware, software, go-to-market capacity, and customer support needed to compete across product categories with very different qualification cycles. A drone platform and an automotive cockpit may both need local intelligence, but they do not share the same safety case, production schedule, field conditions, or tolerance for software changes.

The company also has to manage continuity for customers building on today's platform. SiMa.ai says current applications will be able to migrate to its next-generation architecture. That promise is commercially important because design wins in embedded systems can take years to reach volume, and the software investment made before production often outlives the first piece of hardware.

Commercial Evidence and Its Limits

SiMa.ai says revenue quadrupled between 2024 and 2025 and that momentum continued upward in 2026. The company has not disclosed audited revenue, unit shipments, margins, profitability, or a customer count, so the growth rate is useful directional evidence rather than a complete financial picture.

Its combined customer-and-partner roster includes ARK Electronics, AVerMedia, Bosch, Emerson, Intrinsic, Kontron, L&T Technology Services, Micron, STIGA, Synopsys, TRUMPF SE, and Virya Autonomous Technologies. A joint automotive collaboration with Synopsys shows how SiMa.ai is positioning the platform inside a wider design flow that includes architecture exploration, virtual prototyping, hardware verification, and software development.

The important distinction is between ecosystem access and repeatable production scale. Partners can shorten integration, distribution, and qualification work. They do not, by themselves, prove that a chip platform has won durable volume across several end markets. The Series C gives SiMa.ai the balance sheet to pursue that proof.

Why Physical AI Changes the Chip Sale

Cloud AI rewards scale in centralized infrastructure. Physical AI has to bring enough intelligence into the machine while staying inside the machine's physical limits. A robot, autonomous drone, industrial camera, or vehicle cannot always send every input to a data center and wait for an answer.

That changes the buying criteria. Performance still matters, but performance per watt, local latency, model support, software portability, sensor integration, reliability, and product lifecycle become part of the same decision. The best benchmark cannot rescue a platform that takes too long to integrate or leaves an engineering team trapped on a narrow toolchain.

SiMa.ai's full-stack approach is meant to compete on that broader decision. The company is not simply asking customers to buy an accelerator; it is asking them to trust a software environment, silicon roadmap, migration path, partner network, and support model across products expected to operate for years.

What the Valuation Now Requires

The $1.45B valuation places more weight on execution than on category discovery. SiMa.ai no longer needs to prove that local AI inference matters. The harder task is turning a broad Physical AI thesis into repeatable design wins and production programs across robotics, drones, automotive, and industrial systems.

The next evidence should come from the handoffs that funding announcements cannot settle: developers moving existing applications onto Palette Neat, customers qualifying Modalix inside real products, next-generation hardware reaching its planned schedule, and those products moving from evaluation to volume. SiMa.ai has financed the stack it wants the market to adopt. The company now has to make a difficult migration feel ordinary inside machines that cannot afford ordinary mistakes.