Quadric, Inc. has finalized the second close of its Series C funding round, accelerating the deployment and commercialization of its programmable AI processor platform. Designed to run complex machine learning models directly at the edge, the technology responds to a mounting global demand for localized, secure, and self-sustained artificial intelligence infrastructure.
Capital Influx to Propel Edge AI Across Global Markets
The extension was led by the International Finance Corporation (IFC), the private-sector arm of the World Bank, bringing the total Series C pool to $46 million and lifting the company’s historic fundraising to $90 million.
This capital injection will directly scale Quadric’s technical support, engineering, and go-to-market teams. The expansion aims to satisfy growing demand from current automotive, AI PC, and enterprise computing clients, while fueling a pipeline of new deployments in humanoid robotics, wearable devices, and smart networking infrastructure.
A syndicate of prior backers deepened their commitments in this closing. Seed lead investor Pear VC alongside Uncork Capital and Series C first-close lead BEENEXT all increased their stakes following a year of substantial commercial traction. Additionally, Offline Ventures co-founded by Dave Morin and James Higa joined the cap table as a new institutional investor. The momentum follows a standout fiscal period for Quadric, which saw its product revenue more than triple and the firm cross into profitability following its initial Series C announcement.
Unlocking Localized AI Beyond the Cloud
The involvement of the IFC underscores a strategic interest in bringing processing capabilities to regions and enterprises that remain underserved by massive, centralized cloud datacenters.
“Quadric addresses one of the most important structural gaps in the AI ecosystem today,” said Mohamed Eissa, Chief Investment Officer at IFC. “Powerful AI cannot remain the exclusive domain of hyperscalers if emerging markets are going to close the digital divide. Quadric’s programmable architecture fundamentally changes the economics: SMEs in emerging markets can now deploy AI on devices they own, without the per-token cloud bills that price them out. That productivity gain directly levels the playing field between small businesses in emerging and developed markets. And critically, building this class of efficient, programmable chips creates exactly the kind of high-value semiconductor and AI engineering talent that emerging markets like India need to compete globally.”
Also Read: 1stProtect and Multiverse Computing Forge Strategic Alliance to Champion Secure, Cloud-Independent On-Device AI Inference
Traditional neural processing units (NPUs) frequently struggle to keep pace with algorithmic innovation due to the long development cycles required for silicon hardware design.
“A chip feature set is locked years before it ships, and AI models change every few months, so an operator-centric, fixed-function NPU arrives behind the models and only falls further back,” said Veerbhan Kheterpal, CEO and co-founder of Quadric. “Quadric is a living platform: because the stack is software, the same chip runs new models and gets faster long after it ships. That’s the difference between silicon that depreciates and silicon that compounds. Ask our customers.”
“Every NPU gets judged the day a new model drops,” said Daniel Firu, co-founder and Chief Product Officer of Quadric. “We port new models to Chimera cores and our customers take them as a software update, no silicon change. That porting machine is the product: the same core runs models published years after the silicon was designed.”
“We’re excited to keep backing Quadric as it pursues a massive opportunity. AI is moving outside the datacenter, and chip companies need silicon that can run tomorrow’s models, not just today’s. Quadric is solving that, and the design wins are proving it out. We led Quadric’s seed round and doubled down in this round because, from day one, we’ve seen firsthand how innovative this team is, and how strong they are in both the technology and the execution it takes to win,” said Mar Hershenson, Founding Managing Partner of Pear VC, which led Quadric’s seed round.
Architectural Advantages of Chimera Processors
Key to the platform is the Chimera General Purpose NPU (GPNPU). The system scales from 1 to over 3200 TOPS (Tera Operations Per Second) in multiple chiplet configurations. In contrast to inflexible and single-function accelerators, this architecture is compatible with a variety of mathematical systems like conventional computer vision, convolutional networks, transformers, edge LLMs, and new Vision Language Action (VLA) world models.
The architecture allows silicon designers to incorporate the IP once and continuously update the application logic through software defined execution. Quadric’s native software execution environment enables the compilation of models into C++, as well as providing the ability to write custom code using Python or C++.
