Quadric’s Series C Signals On-Device AI Boom

Quadric’s new Series C round is another strong signal that on-device AI is shifting from “nice to have” into a mainstream infrastructure decision for enterprises, device makers, and governments. Here’s a WordPress-ready, long-form news rewrite tailored for the ai world organisation audience and aligned to ai world organisation events and ai conferences by ai world.

Quadric’s Series C and what it signals

Quadric, a Bay Area chip-IP company focused on enabling on-device AI inference, has announced an oversubscribed $30 million Series C round, taking its total capital raised to $72 million. The round was led by ACCELERATE Fund, managed by BEENEXT Capital Management, with Uncork Capital returning through its opportunity fund and Pear VC also participating, alongside new investors including Volta, Gentree, Wanxiang America, Pivotal, and Silicon Catalyst Ventures.

While funding headlines often focus on the round size, the more important takeaway is the “why now” behind this raise: product revenues more than tripled in 2025 versus 2024, and Quadric is positioning that traction as proof that on-device inference is becoming a priority across edge LLM use cases, automotive, and enterprise vision. In a separate interview, CEO and co-founder Veerbhan Kheterpal said Quadric posted $15 million to $20 million in licensing revenue in 2025, up from around $4 million in 2024, and the company is targeting up to $35 million this year as it expands a royalty-driven model. Kheterpal also said this revenue momentum has supported a higher post-money valuation range of roughly $270 million to $300 million.

For the ai world organisation community, this is more than a startup funding update, because it reflects a deeper market rebalancing: who will own inference, where it will run, and how organizations will manage cost, privacy, latency, and sovereignty at scale. This is exactly the kind of strategic infrastructure shift that shapes agendas at the ai world summit and informs the conversations we curate through ai world organisation events and ai conferences by ai world.

Why on-device AI is accelerating now

One reason Quadric’s story is resonating is that the market narrative has changed from “cloud-first AI” to a more pragmatic split where training may remain centralized, but inference increasingly needs to run close to the user, close to the sensor, or inside the enterprise boundary. TechCrunch framed this shift as companies and governments looking to run AI locally to reduce cloud infrastructure costs and build sovereign capability, rather than sending every query to distant servers.

This transition is not only about saving money, even though cloud costs are becoming a board-level concern when AI features scale from pilot projects to everyday workflows. It is also about performance and reliability, because many high-value AI applications—driver assistance, industrial inspection, operational safety, and real-time decisioning—can’t tolerate unpredictable latency or network dependency. When inference happens locally, devices can respond instantly, and systems can keep working even when connectivity is limited, which is crucial in industrial settings, vehicles, and distributed enterprise environments.

There is also a governance layer pushing the market toward local execution. Kheterpal said Quadric is seeing interest from markets exploring “sovereign AI” strategies aimed at reducing reliance on U.S.-based infrastructure, and that the company is exploring customers in countries including India and Malaysia while shaping an India “sovereign” approach with strategic investor involvement. For AI leaders, “sovereign” does not just mean where compute runs; it often extends to how data is handled, where models are hosted, and how compliance is enforced across jurisdictions.

From the perspective of the ai world organisation, this is a major theme for ai world summit programming: practical adoption meets policy constraints, and architecture decisions begin to look like business strategy rather than pure engineering preference. If your organization is building AI into consumer devices, enterprise hardware, or critical infrastructure, the on-device trend is less about hype cycles and more about operational control, total cost of ownership, and long-term resilience.

Quadric’s programmable IP approach

Quadric’s key proposition is that it does not manufacture chips; instead, it licenses programmable AI processor IP that customers can embed into their own silicon, supported by a software stack and toolchain designed to run models on-device. Kheterpal described this licensing model as providing a “blueprint” that enables customers to build their own chips while still benefiting from a programmable approach to inference.

In its Series C announcement, Quadric positioned itself as “the inference engine that powers on-device AI chips,” and said it is seeing accelerating design-win momentum as adoption grows for its General Purpose NPU (GPNPU) processor IP across edge LLM, automotive, and enterprise vision applications. The company argues that many edge AI chips today rely on older architectures with NPU accelerators added as an afterthought, and that software toolchains are often stitched together in ways that don’t scale when developers try to run new models efficiently. Quadric’s pitch, in contrast, is that the real risk in edge AI is not just performance—it is obsolescence, because model architectures change faster than silicon design cycles.

Quadric says its Chimera processor IP is built for this reality: a fully programmable architecture intended to run current and future AI models on a unified design, reducing the risk of model-driven silicon obsolescence. The company adds that Chimera, combined with its toolchain, is designed to enable chip designers to deploy computer vision and on-device LLM applications, including models up to 30 billion parameters, with strong inference performance per watt. Quadric also claims customers can move from engagement to production-ready, LLM-capable silicon in under six months, which—if consistently achieved—changes the planning horizon for teams building AI-enabled devices.

On the scale and configurability front, Quadric states that Chimera GPNPU cores scale from 1 TOPS to 864 TOPS and are available in both commercial-grade and automotive safety-enhanced (ASIL-ready) configurations. TechCrunch also highlighted that Quadric began in automotive and later expanded toward laptops and industrial devices, reflecting how on-device inference moved from a niche requirement into broader product categories as transformer-based models spread.

For the ai world organisation ecosystem, there is a bigger lesson here: the winners in on-device AI may not be the companies with a single “best” model, but the companies building flexible platforms that survive multiple model generations and multiple deployment environments. This is why AI infrastructure leaders, product owners, and policy stakeholders should treat programmable inference and tooling maturity as first-class evaluation criteria—topics that fit naturally into the ai world summit 2025 / 2026 discussion track.

Traction, customers, and global expansion

Quadric’s growth story is being validated through both revenue indicators and customer signals. In its press release, the company said its product revenues more than tripled in 2025 versus 2024, and tied its Series C to accelerating design wins across edge LLMs, automotive, and enterprise. The announcement also quoted BEENEXT’s Hero Choudhary pointing to Quadric’s architecture, its approach to edge inference, and traction “particularly in Asian markets” as factors supporting the investment thesis.

On the customer side, TechCrunch reported that Quadric’s customers span printers, cars, and AI laptops, and specifically mentioned Kyocera and Japan’s auto supplier Denso, which builds chips for Toyota vehicles. The same report said the first products based on Quadric’s technology are expected to ship this year, beginning with laptops. Quadric also said that alongside the funding, it announced two new license wins: an edge-server LLM silicon provider in Asia (name withheld pending product announcement) and Tier IV of Japan, described as a pioneer in self-driving software.

Geographically, TechCrunch noted that Quadric is based in San Francisco and has an office in Pune, India, while employing nearly 70 people worldwide, including around 40 in the U.S. and about 10 in India. That India connection matters for the ai world organisation audience, because it signals that edge AI capability-building is not only a Silicon Valley trend; it is becoming part of how APAC markets approach sovereignty, device innovation, and enterprise AI rollouts.

At the market-structure level, TechCrunch pointed out a core tension: chip development takes years, while model architectures can change in months, which pushes buyers toward programmable platforms that can evolve through software updates instead of repeated hardware redesigns. Kheterpal positioned Quadric as a programmable alternative that helps customers support new models via software updates rather than redesigning silicon each time the dominant architecture changes. This aligns with Quadric’s own framing that customers can’t afford to bet on an architecture that becomes obsolete as models shift.

For the ai world organisation, these details connect directly to the purpose of ai world organisation events: we bring together the people making these long-cycle decisions—founders, enterprises, policymakers, and builders—so they can validate where the industry is heading before committing budgets and roadmaps. It also helps explain why “distributed AI” is no longer an abstract term; it is becoming a practical deployment pattern where inference runs on laptops, on-premise servers, and specialized edge hardware inside offices and industrial sites.

What this means for the AI World community

The AI World Organisation describes its mission as bridging the gap between cutting-edge AI innovation and real-world application while building a global ecosystem through collaboration. Developments like Quadric’s Series C are relevant because they sit exactly at that bridge: the innovation is in programmable inference IP and tooling, and the real-world application is in how enterprises and governments actually deploy AI under cost, privacy, and sovereignty constraints.

For members, delegates, and partners engaging with the ai world summit, the most productive way to read this story is not “a chip IP startup raised money,” but “on-device inference is becoming a mainstream architectural choice, and platform bets are shifting accordingly.” The implications show up across multiple tracks that the ai world organisation and ai world organisation events regularly cover: enterprise AI operating models, public-sector AI readiness, privacy-by-design deployments, and the economics of inference at scale.

This is also where the event layer matters. The AI World Organisation’s upcoming global summits list includes multiple 2026 events across regions, including AI World Summit 2026 Asia in Singapore (28th May, 2026) and additional “register interest” listings for Dubai, Sydney, Amsterdam, and London later in 2026, alongside India-based events such as the GCC Conclave (14th March, 2026, Hyderabad) and the Talent, Tech & GCC Summit (17th April, 2026, Delhi). If your organization is deciding whether inference should live in the cloud, on-device, or in a hybrid distributed model, these are precisely the rooms where you can compare playbooks across industries and regions—making the ai world summit 2025 / 2026 cycle especially timely for infrastructure and product leaders.

For context, The AI World Summit 2025 took place on 17–18 January 2025 and was positioned as a gathering of AI leaders, innovators, and professionals across policy, research, and industry. As the next wave of ai world organisation events unfolds, the on-device AI story—represented by companies like Quadric—will likely remain front and center because it intersects with sovereignty, cost control, and practical deployment realities in enterprise and public sector AI.

In short, Quadric’s rise is a case study in how the AI market is maturing: inference is moving closer to users, programmability is becoming insurance against rapid model change, and the strongest momentum is appearing where engineering constraints and policy priorities meet. For the ai world organisation audience, this is exactly the kind of signal worth tracking, debating, and translating into execution frameworks at the ai world summit, and across ai conferences by ai world in 2026.