Quadric Extends Series C to $46M for On-Device AI

The AI infrastructure conversation has become predictable: bigger models, bigger data centers, and taller GPU stacks. Quadric, a Burlingame, California semiconductor IP company developing programmable AI processor IP for on-device AI inference, operates in the quieter layer beneath that conversation. On July 13, 2026, the company announced it had extended its Series C financing to $46M, bringing total funding to $90M. The extension was led by the International Finance Corporation, with participation from Pear VC, Uncork Capital, BEENEXT, and Offline Ventures.

Led by CEO Veer Kheterpal, CTO Nigel Drego, and CPO Daniel Firu, Quadric is not trying to manufacture chips itself. Instead, it licenses processor IP to semiconductor companies building custom silicon, giving customers a programmable architecture for AI inference workloads that run closer to devices, machines, and users rather than relying entirely on centralized cloud infrastructure.

What Happened

Quadric's latest Series C extension brings a new lead investor into a round that already included strong support from existing backers. The extension was led by the International Finance Corporation, with participation from Pear VC, Uncork Capital, BEENEXT, and Offline Ventures. The investor mix matters because it combines venture continuity with an institution whose mandate extends well beyond traditional Silicon Valley investing.

The company develops and licenses Chimera GPNPU, its General Purpose Neural Processing Unit architecture. GPNPU stands for General Purpose Neural Processing Unit, a programmable processor architecture designed for AI inference workloads. In practical terms, Quadric is providing the technology that allows chip designers to build AI-capable silicon without requiring every new model update to trigger a hardware redesign.

Why This Matters

AI infrastructure is beginning to split into two worlds: massive centralized compute and distributed intelligence running closer to where decisions happen. Quadric's Chimera GPNPU belongs to the second world, where vehicles, industrial systems, enterprise equipment, AI PCs, robotics, networking hardware, and edge devices need inference capabilities without relying entirely on remote data centers. That makes programmable on-device AI less of a feature and more of a systems requirement.

The market is also entering a different phase. Training frontier models continues to dominate headlines, but deployment is where much of the commercial friction exists. If AI workloads continue evolving faster than hardware cycles, programmable processor IP gives semiconductor companies a way to preserve flexibility inside products expected to remain in service for years.

Market Context

Quadric's business model positions the company inside the semiconductor ecosystem rather than the application layer. It licenses processor IP instead of fabricating chips, meaning its success depends on convincing chipmakers and hardware companies that programmable AI inference belongs in their silicon roadmaps. That is a more complex sale than a software subscription, but it can become deeply embedded once customers commit.

The company also reported meaningful commercial momentum alongside the financing. Quadric said product revenue more than tripled in 2025 compared with 2024 and that it reached profitability before extending the Series C round. Those details make the financing look less like speculative AI enthusiasm and more like capital following execution as the market becomes increasingly focused on where AI inference actually happens.

What This Signals

One of the strongest signals in venture capital is not who writes the first check. It is who continues investing after seeing the operating reality up close. Pear VC, Uncork Capital, and BEENEXT returning to the round suggests confidence built through execution rather than another polished funding narrative.

IFC's participation sends a different signal. As the private-sector arm of the World Bank Group, it often backs infrastructure with the potential to matter across regions, industries, and broader development priorities. Its investment suggests on-device AI infrastructure is being viewed not only as a venture category but also as foundational technology for expanding AI capabilities in markets where centralized cloud infrastructure may be expensive, constrained, or strategically undesirable.

The Bigger Industry Shift

The industry will continue celebrating larger models and larger compute clusters because those stories are easy to see. The quieter transformation is happening inside the processor architectures that make AI practical outside the data center. Quadric's latest financing places the company within that shift, alongside the chip designers and hardware builders working to make intelligence more distributed, programmable, and economically practical.

That is the operator lesson behind the round. Markets do not reward only the companies with the flashiest AI demonstrations. They also reward the companies removing structural friction for the builders underneath those demonstrations. Quadric's bet is that as AI moves from spectacle to infrastructure, the winners will be the companies that make the next generation of workloads easier to run where the work actually happens.