Skip to main content

Quadric

Category: AI Infrastructure

A developer of the Chimera GPNPU (General Purpose Neural Processing Unit), a unified processor IP architecture that combines the high-performance efficiency of an NPU with the full programmability of a CPU and DSP. Quadric was founded in 2016. The company is led by Veerbhan Kheterpal. Based in Burlingame, USA. Team size: 50-100. Total funding raised: $72.0M ($42M + $30M Series C). Latest round: Series C ($30M, Jan 2026). Key investors include BEENEXT, NSITEXE (Denso), MegaChips, Uncork Capital, Pear VC, Volta Ventures.

Founded
2016
Headquarters
Burlingame, USA
Team size
50-100
Total funding
$72.0M ($42M + $30M Series C)

Value proposition

Eliminates the 'two-processor' bottleneck by allowing developers to run both neural network inference and general-purpose C++ code on a single unified core, significantly reducing latency, power consumption, and software complexity.

Products and solutions

Chimera GPNPU IP Cores (QB Series and QC Series), Chimera QC-M (Multi-core cluster scaling up to 864 TOPs), Chimera SDK / Developers Studio (Unified C++ programming environment), Chimera Graph Compiler (Automated model conversion), Safety Enhanced (SE) Versions (ASIL-B and ASIL-D compliant for automotive)

Unique value

Pioneered the 'GPNPU' category, which treats AI as a general-purpose computing task rather than a fixed-function 'black box' acceleration, enabling the entire application pipeline to run on one core.

Target customer

System-on-Chip (SoC) developers, semiconductor companies, and OEMs in the automotive, industrial, and consumer electronics sectors.

Industries served

Automotive (ADAS and Autonomous Driving), Industrial Robotics and Automation, Consumer Electronics (IoT, Wearables, AI PCs), Medical Imaging, Enterprise Vision Systems

Technology advantage

Provides a single-source C++ programming model that removes the need for manual code partitioning between NPUs and DSPs, offering up to 2.7x greater compute density (TOPs/mm2) than previous generations.

How they differentiate

Quadric differentiates through its 'GPNPU' (General Purpose Neural Processing Unit) architecture. Unlike traditional NPUs that require a separate DSP or CPU to handle non-neural network code (pre/post-processing), Quadric’s Chimera architecture runs both AI inference and general-purpose C++ code on a single unified core, eliminating data movement bottlenecks and simplifying software development.

Main competitors

Arm (Ethos-U and Ethos-N series), Cadence (Tensilica AI IP), Synopsys (DesignWare ARC EV Processors), Hailo, Blaize

Key partnerships

DENSO (Co-development of in-vehicle AI semiconductors), Kyocera (Licensing partner for consumer and automotive products), MegaChips (Strategic investor and integration partner), NSITEXE (Subsidiary of DENSO for automotive IP), Tier IV (Autonomous driving software integration)

Notable customers

Denso, Kyocera, MegaChips, Tier IV

Major milestones

Pioneered the 'GPNPU' category for unified AI and C++ processing., Achieved ASIL-B and ASIL-D safety compliance for automotive applications., Successfully transitioned from a hardware-chip company to a Silicon IP licensing model., Closed $30M Series C in January 2025 to fuel global expansion.

Growth metrics

Scaled licensing revenue by 3x in 2024; projected revenue of $15M+ for 2025.

Market positioning

High-performance Silicon IP provider for Edge AI, targeting the 'middle ground' between low-power microcontrollers and high-end server GPUs.

Geographic focus

Global, with a primary focus on North America and the Asia-Pacific region (specifically Japan, due to strategic partnerships with Denso and MegaChips).

Patents and IP

Holds a significant portfolio of patents related to parallel processing architectures, data flow management, and unified memory structures for on-device AI (specific public filings include proprietary GPNPU architecture designs).

About Veerbhan Kheterpal

Veerbhan Kheterpal is a serial entrepreneur and semiconductor expert with over 20 years of experience in silicon and software. He was a technical co-founder of 21.co (formerly 21inc, acquired by Coinbase), where he led the development of power-efficient ASICs. Previously, he served as VP of Engineering at Factual (acquired by Foursquare) and held senior engineering roles at PDF Solutions. He holds a PhD and MS in Electrical and Computer Engineering from Carnegie Mellon University and a B.Tech from IIT Kharagpur.

Official website: