
OXMIQ Labs, Inc., a Campbell, California-based developer of unified GPU and AI chiplet architectures, has raised $35 million in a Series A funding round led by Fundomo and Samsung Catalyst Fund.
The round also saw participation from MediaTek, AM Intelligence Labs, Pegatron Venture Capital, CDIB-TEN, Darwin Ventures, and Morgan Creek Digital.
The company will use the funds to increase production of its GPU technology, expand its licensing business, accelerate the adoption of its chiplet integration platform, and grow its hardware and software engineering teams that support customers.
OXMIQ develops technology across the entire AI stack, including renewable energy, data center infrastructure, silicon IP, electron-to-token machines (ETMs™), and the software that powers AI factories and AI agents.
RECOMMENDED FOR YOU
[Funding news] Canada-based Nulogy Secures $15Million in Credit Facilities
Startuprise io
Sep 27, 2024
Read More:Addi Raises $85M in Series D Funding Led by BTG Pactual
As demand for AI computing continues to grow faster than available infrastructure, OXMIQ is redesigning the GPU stack from "Atoms to Agents™." The company builds silicon IP, configurable hardware systems, and software platforms that help semiconductor companies and AI infrastructure providers reduce the cost of AI computing.
At the core of OXMIQ's technology is OxCore™, a scalable and licensable GPU architecture that combines three compute engines into a single design: a CUDA®-compatible GPU engine, a tensor processing engine, and a CPU-based orchestration engine that manages AI workloads and agents.
By combining functions that are typically spread across multiple chips, OxCore reduces data movement and improves performance and energy efficiency for AI applications. The architecture is designed to scale from small AI systems to large data centers. OxCore is currently running on FPGA hardware and is available for live demonstrations.
OxQuilt™ is OXMIQ's chiplet integration platform that combines different compute chiplets and memory into a single package. Unlike many AI chip designs that depend on a specific foundry or memory type, OxQuilt supports multiple process nodes, memory technologies, interconnect standards, and packaging options.
This flexible approach allows companies to build custom AI chip's without the cost and complexity of developing a full chip from scratch. The platform is also designed to support future technologies, including silicon photonics as they become commercially available.
OXMIQ complements its hardware with a software platform that includes OxCapsule™ for AI workload management and OxPython™ for running existing CUDA® and PyTorch® applications without changing the code. This allows developers to move their AI workloads easily across different hardware while supporting new AI models and high-performance inference. OxPython has already been tested on third-party platforms and is available for live demonstrations.
The company follows an IP first business model, focusing on developing AI architecture and licensing its technology instead of building complete chips. This approach helps generate revenue through customer partnership's while keeping development costs lower.
“I am excited to join the OXMIQ board. Raja and this team are creating an open GPU architecture, a much-needed step toward removing the artificial boundaries around AI innovation. As the industry concentrates around a few incumbents, this is more important than ever. OXMIQ’s open, configurable foundation, which developers can build on and own, is exactly where compute should be heading,” said Jim Keller, CEO of Tenstorrent and OXMIQ board member.
“We are very excited to co-lead OXMIQ’s financing round and back Raja Koduri and the strong team at OXMIQ,” said David (Dede) Goldschmidt, SVP & Managing Director, Head of the Samsung Catalyst Fund. “OXMIQ’s novel AI core and software platform enable heterogeneous compute for efficient, custom inference solutions serving large-scale agentic workloads.”
“Raja has built silicon at every layer of the stack, and he knows exactly where the constraints sit. Most compute IP makes the customer bend their memory, packaging, and foundry around the chip. OXMIQ does the opposite, and that flips a cost center into leverage. We backed this team because they will define how AI compute gets built this decade,” said Rajeev Surati, Partner at Fundomo.
About OXMIQ Labs
Founded in 2023 by Raja Koduri, OXMIQ Labs is an AI and GPU architecture company that develops technology across the AI stack, including silicon IP, AI hardware, data center infrastructure, and cloud software. The company's products—OxCore™, OxQuilt™, OxPython™, and OxCapsule™—help semiconductor companies, cloud providers, and AI system builders develop, customize, and deploy AI computing solutions for a wide range of applications.
Read More:10Beauty Raises $23.5M in Funding Led by Story Ventures
Recommended Stories for You
Arsenal Capital Partners Adds Max Schechter as Head of Industrial Growth Business Development
Startuprise io Jun 20, 2026







