Oxmiq Raises $35M for AI Chip Architecture Licensing

Oxmiq, the AI chip architecture startup founded by industry veteran Raja Koduri, has successfully closed a $35 million funding round. This infusion of capital aims to accelerate the development and market introduction of a novel, licensable chip architecture designed to significantly reduce the cost and complexity of custom AI silicon. The company has now raised a total of $60 million, underscoring strong investor confidence in its vision to democratize AI hardware development.

The latest financing was co-led by Samsung Catalyst Fund and Fudomo, with participation from notable investors including MediaTek and Pegatron Venture Capital. This strategic backing positions Oxmiq to challenge the status quo in AI chip design, where the creation of bespoke processors has historically demanded substantial upfront investment and lengthy development cycles, often running into hundreds of millions of dollars and years of engineering effort.

At its core, Oxmiq is developing a foundational IP, dubbed OxCore™, intended to serve as a blueprint for semiconductor manufacturers and AI system integrators. This licensable GPU architecture is engineered to enable the creation of custom AI processors without requiring companies to undertake the full, resource-intensive process of designing a chip from the ground up. This approach is particularly relevant in a market experiencing escalating demand for specialized AI hardware across cloud infrastructure, enterprise solutions, and AI model development.

The market for custom AI silicon is rapidly expanding, driven by hyperscalers and enterprises seeking to optimize performance and reduce reliance on general-purpose computing units. Companies like Broadcom and Marvell are already active in this space, but Oxmiq aims to differentiate by offering a more accessible architectural foundation. This strategy mirrors the impact Arm had on the mobile industry, where its licensed designs became ubiquitous, fostering innovation and widespread adoption by enabling numerous companies to build their own devices.

Raja Koduri, a former key architect at Intel and AMD, envisions Oxmiq becoming the foundational IP provider for the next generation of AI hardware, much like Arm did for mobile processors. The company's roadmap extends beyond core processor designs to include advanced computing fabrics that integrate chiplets and memory, addressing the industry's shift away from monolithic chip designs towards more modular and efficient architectures.

With the new funding, Oxmiq plans to finalize its initial architecture and software platform, expand its engineering talent pool, and prepare for market entry. The company's focus on heterogeneous compute for efficient, custom inference solutions is seen as a critical enabler for large-scale agentic workloads, a growing segment within AI applications. This strategic move addresses the increasing need for specialized hardware as AI models continue to grow in complexity and inference demands surge.