Quadric Triples Licensing Revenue and Raises $30M to Scale On-Device AI IP Business

New updates have been reported about Quadric.

Quadric is emerging as a key beneficiary of the shift from cloud-based AI to local, on-device inference, reporting a sharp acceleration in both revenue and financing as customers seek to cut cloud costs and build sovereign AI capabilities. The San Francisco–based chip IP company, founded by veterans of early bitcoin mining firm 21E6, grew licensing revenue to an estimated $15 million–$20 million in 2025 from roughly $4 million in 2024, according to CEO Veerbhan Kheterpal, and is targeting up to $35 million this year as it transitions toward a royalty-driven model. That growth has lifted Quadric’s post-money valuation to about $270 million–$300 million, up from around $100 million at its 2022 Series B, and underpinned a new $30 million Series C led by ACCELERATE Fund, managed by BEENEXT Capital Management, bringing total funding to $72 million. Quadric does not manufacture chips; instead, it licenses programmable AI processor IP—described as a “blueprint” plus software stack and toolchain—that customers integrate into their own silicon to run vision, voice and other AI models locally on devices such as printers, cars and AI laptops, with initial products expected to ship this year, starting in laptops.

Strategically, Quadric is positioning itself as the on-device counterpart to data-center platforms like Nvidia’s CUDA, arguing that customers need programmable processor IP that can adapt to rapidly changing AI models through software updates rather than costly hardware redesigns. The company initially focused on automotive, enabling real-time functions such as driver assistance, but the rapid adoption of transformer-based models since 2023 has expanded demand across broader device categories as enterprises and governments shift inference from centralized hyperscale infrastructure to distributed setups on laptops and small on-premise servers. Quadric now counts customers including Kyocera and Denso, which supplies chips to Toyota, and is actively pursuing “sovereign AI” opportunities in markets like India and Malaysia, where policymakers and corporates want domestic AI compute capabilities and reduced dependence on U.S.-based cloud providers. Kheterpal positions Quadric’s IP as an alternative to vertically integrated chip vendors such as Qualcomm, whose AI capabilities are tied to their own processors, and to traditional IP suppliers like Synopsys and Cadence, which he argues offer harder-to-program neural processing blocks. Despite employing only about 70 people globally, with the majority in the U.S. and a growing presence in India, Quadric’s longer-term upside hinges on converting today’s licensing deals into high-volume deployments that generate recurring royalties as the industry’s move toward distributed, sovereign and cost-efficient AI inference accelerates.