Quadric moves from cloud AI to device intelligence — and it’s paying off

Companies and governments are looking for tools to manage AI locally in aa bid slash cloud infrastructure investment and independent power generation. QuadricChip-IP, a startup founded by veteran bitcoin mining firm 21E6, is trying to fuel the revolution, expanding beyond cars to laptops and industrial devices, with its device-targeting technology.

That growth is already paying off.

Quadric has posted $15 million to $20 million in returns in 2025, up from $4 million in 2024, CEO Veerbhan Kheterpal (pictured above, center) told TechCrunch in an interview. The company, which is based in San Francisco and has an office in Pune, India, is seeking $35 million in funding this year as it builds an AI business powered by royalty. That growth has boosted the company, now valued at between $270 million and $300 million, from its $100 million Series B round in 2022, Kheterpal said.

It has also helped attract investors to the company. Quadric he announced last week’s $30 million Series C round led by ACCELERATE Fund, led by BENEXT Capital Management, bringing its total funding to $72 million. The upgrade comes as investors and chip makers look for ways to push more AI from the central cloud to local devices and servers, Kheterpal told TechCrunch.

From cars to everything

Quadric started in carswhile on-device AI can perform real-time tasks like driver assistance. Kheterpal said the proliferation of transformer-powered models in 2023 had made “everything” more difficult in the past 18 months as more companies tried to run AI locally instead of relying on the cloud.

“Nvidia is a strong data AI platform,” said Kheterpal. “We wanted to create a CUDA-like toolkit or an alternative to AI tools.”

Unlike Nvidia, Quadric does not make its own chips. Instead, it licenses the AI ​​processor IP, which Kheterpal described as a “system” that customers can embed in their own silicon, along with software and modeling tools, including vision and voice, on devices.

Initial customers include printers, cars, and AI laptops, including Kyocera and Japanese automaker Denso, which makes chips for Toyota cars. The first products based on Quadric technology are expected to ship this year, starting with laptops, Kheterpal told TechCrunch.

Even so, Quadric is now looking beyond traditional commercial services and into markets looking for “AI” solutions to reduce reliance on U.S.-based infrastructure, Kheterpal said. The startup is targeting customers in India and Malaysia, he added, counting Moglix CEO Rahul Garg as an entrepreneur who helps shape its India strategy “independently”. Quadric employs about 70 people worldwide, including about 40 in the US and about 10 in India.

The push is being driven by the rising cost of AI-centric infrastructure and the challenges many countries face in building data centers, Kheterpal said, which is driving more interest in “AI” implementations that focus on laptops or small servers inside offices rather than relying on cloud services for every query.

The World Economic Forum to point for this change in the current context, as AI concepts move closer to users and away from centralized architecture. Similarly, EY he said in a November report that an autonomous AI approach is gaining momentum as policymakers and industry groups push for in-house AI capabilities that combine compute, modeling, and data, rather than relying on external tools.

For chipmakers, the problem is that AI models are developing faster than hardware design, Kheterpal said. He added that customers need a stable IP processor that can keep up with software changes rather than requiring costly upgrades every time a building changes from older vision-based systems to today’s flexible systems.

Quadric is positioning itself as an alternative to chip vendors such as Qualcomm, which often uses its AI technology inside its processors, and IP providers such as Synopsys and Cadence, which sell neural processing engine blocks. Kheterpal said Qualcomm’s approach can lock customers into their own silicon, while traditional IP vendors offer engine blocks that many customers find difficult to implement.

Quadric’s modular approach allows customers to support new AI models through software updates instead of hardware upgrades, providing an advantage in industries where chip development can take years, while Model architecture changes over the course of months these days.

However, Quadric is still in the early stages of its construction process, with few customers signed up to date and its long-term sustainability dependent on converting today’s licenses to deliver high volume and recurring revenue.