Revolutionizing AI: Quadric’s Leap into On-Device Technology
As businesses and governments seek to reduce cloud infrastructure costs while enhancing local capabilities, the demand for local AI solutions has never been higher. Quadric, a San Francisco-based chip-IP startup founded by veterans of the bitcoin mining firm 21E6, is at the forefront of this transformation. The company is expanding its reach beyond automotive applications into laptops and industrial devices, offering cutting-edge on-device inference technology that aims to bring AI processing closer to users.
Quadric’s efforts are already yielding impressive returns. CEO Veerbhan Kheterpal recently informed TechCrunch that the company generated between $15 million and $20 million in licensing revenue in 2025, a remarkable increase from just around $4 million in 2024. With ambitious goals for up to $35 million in revenue this year, Quadric’s valuation has surged to between $270 million and $300 million—an impressive rise from about $100 million during its 2022 Series B funding round.
Last week, Quadric announced a successful $30 million Series C funding round, led by ACCELERATE Fund managed by BEENEXT Capital Management, pushing its total funding to $72 million. This follows a broader trend where investors and chipmakers are looking to decentralize AI workloads from cloud-based systems to local devices and servers.
From Automotive Roots to Diverse Applications
Initially focused on the automotive sector, where on-device AI facilitates real-time functions like driver assistance, Quadric is now broadening its scope. Kheterpal noted that the rise of transformer-based models in 2023 significantly changed the landscape, pushing businesses to adopt local AI solutions instead of relying solely on cloud services.
He emphasized, “Nvidia is a strong platform for data-center AI. We aim to create a similar programmable infrastructure for on-device AI.” Unlike Nvidia, however, Quadric does not manufacture its chips. Rather, it offers licensed programmable AI processor IP that acts as a “blueprint” for customers to integrate into their own silicon, accompanied by a robust software stack to run various models, including both vision and voice technologies.
The company’s diverse client base includes sectors from printer manufacturers to automotive suppliers like Denso, which provides chips for Toyota. Kheterpal is optimistic, stating that the first products using Quadric’s technology are expected to launch within this year, starting with AI-equipped laptops.
Quadric is also exploring opportunities in “sovereign AI” strategies, aiming to reduce dependence on U.S.-based infrastructure. The startup is eyeing markets in India and Malaysia, benefiting from strategic insights from Moglix CEO Rahul Garg, who is assisting in Quadric’s sovereign initiatives. Currently, Quadric employs nearly 70 people globally, with a workforce split between the U.S. and India.
The rising expenses associated with centralized AI infrastructure, coupled with the challenges some countries face in establishing large-scale data centers, are driving this interest in distributed AI. Kheterpal remarked that more businesses are considering setups where inferences run on laptops or small on-site servers instead of relying entirely on cloud services.
The World Economic Forum highlights this shift, noting a trend towards localized AI solutions. Additionally, a November report by EY underscored the growing traction of sovereign AI, as policymakers and industry players advocate for domestic capabilities spanning computing, data, and models.
For chip manufacturers, the challenge lies in keeping pace with evolving AI models, which outstrip hardware design cycles. As Kheterpal explained, there’s a pressing need for programmable processor IP that can adapt to rapid changes through software updates—avoiding costly redesigns with each architectural evolution from earlier models to today’s transformer architectures.
Quadric is positioning itself as an alternative to established chip vendors like Qualcomm, which generally embeds its AI tech into proprietary processors. Kheterpal advocates that this approach can potentially lock customers into a single silicon supplier. In contrast, Quadric’s programmable model allows customers flexibility, facilitating easier updates to new AI models without necessitating a hardware overhaul.
While Quadric has made significant strides, the company acknowledges that it is still in the early stages of its journey. The ongoing success will depend on converting current licensing deals into high-volume shipments and establishing a steady stream of recurring royalties.
For more insights into emerging tech trends and investment opportunities, visit AxomLive.
