OXMIQ Raises $35 Million to Scale OxCoreTM Architecture

OXMIQ, a GPU and AI architecture company with expertise spanning every layer of the AI stack, led by Raja Koduri and the team, has closed its $35 million Series A financing, bringing the company's total capital raised to $60 million. The round was co-led by Fundomo and Samsung Catalyst Fund, with participation from MediaTek, AM Intelligence Labs, Pegatron Venture Capital, CDIB-TEN, Darwin Ventures, and Morgan Creek Digital, among other financial and strategic investors.

OXMIQ’s expertise spans the full AI stack, from renewable power and data center infrastructure to silicon IP, electron-to-token machines (ETMs™), and the software powering AI factories and agents.

One Core, Three Engines.

Token demand is outpacing global infrastructure capacity. OXMIQ was founded to re-architect the GPU stack from Atoms to Agents™, building silicon IP, configurable systems, and software that reduce the cost of intelligence across the stack.

At the center of the architecture is OxCore™, a scalable, licensable GPU core that integrates three compute engines: a CUDA®-compatible GPU engine, a tensor processing engine, and a CPU-based orchestration engine. Designed for near-memory compute, OxCore minimizes data movement while improving efficiency across AI workloads. The architecture scales from single-core deployments to large datacenter systems, and is currently running on an FPGA with live demonstrations.

OxQuilt™, the chiplet integration layer, combines heterogeneous compute and memory into a unified package. Instead of being locked to a single foundry or memory configuration, OxQuilt enables flexible system design across process nodes, interconnects, and packaging technologies, including emerging silicon photonics.

The hardware stack is paired with a software layer spanning OxCapsule™ for orchestration and OxPython™ for developer compatibility. OxPython enables CUDA® and PyTorch® workloads to run without code changes, ensuring portability and day-zero support for new models.

Together, this IP-first approach enables capital-efficient development by monetizing architecture and system design rather than full chip fabrication.

“We are very excited to co-lead OXMIQ’s financing round and back Raja Koduri and the strong team at OXMIQ. OXMIQ’s AI core and software platform enable heterogeneous compute for efficient, custom inference solutions serving large-scale agentic workloads,” said David (Dede) Goldschmidt, SVP & Managing Director, Head of Samsung Catalyst Fund.

“Raja has built silicon across every layer of the stack and understands where the real constraints sit. Most compute IP forces customers to adapt memory, packaging, and foundry choices around the chip. OXMIQ takes the opposite approach, turning that constraint into leverage. This team is positioned to define how AI compute is built this decade,” said Rajeev Surati, Partner at Fundomo.

An Expanding Team.

OXMIQ has strengthened its board and advisory group with two additions bringing decades of silicon expertise. Jim Keller, CEO of Tenstorrent and one of the industry’s most influential chip architects, joins the board alongside existing board member Dr. Ker Zhang. Dr. Valluri (Bob) Rao, former Intel Fellow from the process technology group, joins as an advisor. Together, they reinforce OXMIQ’s transition from architecture development to customer integration.

“An open GPU architecture is a necessary step toward removing artificial boundaries around AI innovation. As the industry concentrates around a few incumbents, OXMIQ’s open and configurable foundation allows developers to build and own their systems,” said Jim Keller.

“A licensable core with an open architecture enables design teams to build custom AI silicon tailored to their needs. Reducing compute cost expands access to AI development. The goal is to make AI a tool available to many, not a capability reserved for a few,” said Raja Koduri, Founder & CEO of OXMIQ Labs.

Investors.

The round was co-led by Fundomo, a New York venture firm focused on frontier compute infrastructure, and Samsung Catalyst Fund, Samsung Electronics’ evergreen venture capital arm investing in deep tech AI infrastructure. MediaTek, an early seed investor and leading fabless semiconductor company, is reinvesting in this round.

Additional participation includes AM Intelligence Labs (part of AM Green Group), extending its collaboration on the 5GW AI factory initiative, including a 3GW renewable-powered AI compute platform under development in India. CDIB-TEN, a joint fund combining CDIB Capital’s multi-decade investment platform with TEN Capital’s semiconductor-focused expertise, also joins the round.

Pegatron Venture Capital contributes manufacturing and systems integration expertise, supporting the path from chiplet architecture to deployment. Morgan Creek Digital invests on the thesis that compute architecture and capacity will define the next phase of the AI economy. Darwin Venture Management adds cross-border coverage across the Taiwan–Silicon Valley technology corridor. Intel Capital participates as a strategic investor and IP partner, strengthening the company’s engineering and ecosystem depth.