Raja Koduri's Oxmiq raises $35M to license AI chip designs instead of selling finished chips

Raja Koduri's Oxmiq raises $35m to rent out AI chip design instead of selling chips

OxCore is a licensable GPU core meant to let semiconductor firms skip years of design work. Jim Keller just joined the board to back the idea. Oxmiq Labs has closed a $35m Series A to scale OxCore, a licensable GPU architecture the startup says lets chipmakers build custom AI silicon without running a full, multi-year design programme of their own. The round brings the company's total capital raised to $60m since its founding by veteran chip architect Raja Koduri. The pitch is straightforward even if the engineering underneath it is not. Building an advanced AI chip from scratch typically costs hundreds of millions of dollars and takes years, a barrier that has kept custom silicon out of reach for all but the largest cloud providers and chipmakers. Oxmiq wants to sell the design itself as licensable IP, the way Arm licenses processor cores, rather than selling finished chips. OxCore is the product at the centre of that bet. It integrates three compute engines, a CUDA-compatible GPU engine, a tensor processing engine, and an orchestration engine that coordinates workloads across the system, functions typically split across separate chips. Oxmiq says the tighter coupling is built for near-memory compute, reducing the data movement that drives up both cost and energy use in AI workloads. The round was co-led by Fundomo and Samsung Catalyst Fund, with participation from MediaTek, Pegatron Venture Capital, Darwin Ventures, Morgan Creek Digital, and other strategic and financial investors named in the announcement. That mix, spanning a memory and foundry-adjacent giant in Samsung, a contract manufacturer in Pegatron, and a fabless chipmaker in MediaTek, suggests interest that goes beyond pure financial return. Perhaps the more interesting signal is who just joined the board. Jim Keller, the chip architect whose career runs through Apple, AMD's Zen architecture and Tesla's self-driving silicon before he became chief executive of Tenstorrent, has taken a board seat at Oxmiq. Former Intel process technology fellow Valluri "Bob" Rao has joined as an advisor. Neither appointment is disclosed as carrying an equity or compensation figure. Koduri's own background gives the pitch some weight. He led graphics at AMD, then Apple, then spent years running Intel's graphics and accelerated-computing efforts before striking out on his own, putting him among a small group of architects who have shipped GPU silicon at three of the industry's largest companies. Oxmiq is headquartered in Campbell, California, with a development site in Hyderabad, India. The licensable-IP model itself is not new in chip design. Arm has run a version of it for decades, and RISC-V startups such as SiFive have built entire businesses on licensing processor cores rather than shipping finished silicon. What is less tested is applying that model to frontier AI accelerators, where the biggest buyers, Nvidia chief among them, have historically preferred to own their architecture outright rather than license someone else's. Oxmiq's bet is that enough semiconductor companies and system builders would rather pay for a proven, working core than gamble hundreds of millions of dollars and several years of engineering time building one from nothing. That calculation has become more attractive as AI infrastructure spending has ballooned and more companies want their own custom silicon rather than relying entirely on Nvidia or a hyperscaler's in-house chip programme. The financing terms beyond the headline number were not disclosed. Oxmiq did not specify what stake its new investors took or what near-term revenue, if any, the company is generating from licensing deals already in place, leaving the commercial traction underlying the round largely unverified from the public record. Whether the bet pays off will depend less on the funding round than on whether any of Oxmiq's backers actually put OxCore into silicon that ships. For now, the $35m buys the company runway and a board seat from one of the industry's more credible names, which is its own kind of validation in a sector where trust in an unproven architecture is the scarcest resource of all.

Raja Koduri's Oxmiq Labs raises $35M to lower the design cost of custom AI silicon

Raja Koduri's Oxmiq Labs raises $35M to lower the design cost of custom AI silicon Artificial intelligence chipmaking startup Oxmiq Labs Inc. says it wants to become the next Arm Holdings Plc. after raising $35 million in Series A funding, bringing its total amount raised to date to $60 million. Today's round was led by investors including Fundomo and Samsung Catalyst Fund, and saw participation from MediaTek, AM Intelligence Labs, Pegatron Venture Capital, CDIB-TEN, Darwin Ventures and Morgan Creek Digital. Oxmiq is founded and led by the respected graphics processing unit architect Raja Koduri (pictured), who formerly served as the chief architect and Executive Vice President of Intel Corp.'s architecture, graphics and software division. In an interview with Reuters, Joduri explained that his company is working on a new chip design architecture and software that will dramatically lower the cost of developing AI chips. The funds from today's round will help to scale Oxmiq's proprietary GPU architecture OxCore, which makes it easier for semiconductor firms and AI systems builders to design custom silicon without a full chip program. ' According to Koduri, developing a cutting-edge AI chip usually costs hundreds of millions of dollars and requires several years. It's an intricate process that involves making the silicon design plans and building the software needed to support it. Oxmiq aims to simplify that process by collapsing three distinct components of an AI system into a single intellectual property block it can license to customers. "A licensable core with an open architecture means design teams everywhere can build the custom AI silicon their work needs," Koduri said. Generally, AI chip systems incorporate both GPUs and central processing units, but Oxmiq's plan is to combine them both with a third component it calls a "tensor engine" into a single design, Koduri explained. It will then license the whole thing. "We would want to be the Arm of this next era," Koduri said, in reference to the U.K.-based semiconductor firm that supplies the designs and IP for almost every smartphone chip in the world. Oxmiq is also developing a computing fabric called OxQuilt, which combines heterogeneous compute chiplets and memory into a single package. Chiplets are several specific chips that form a complete system. With OxQuilt, chip designers will be able to adapt to any supply chain using configuration tools that let customers mix and match different process nodes, memory types, interconnect standards and advanced packaging systems. The architecture is also designed to incorporate emerging interconnects based on silicon photonics when they reach maturity. Meanwhile, its software stack includes something called OxCapsule for orchestration and OxPython, which makes it possible to run existing CUDA and PyTorch code on OxCore without modifying it first. Koduri said Oxmiq eventually wants to compete in the custom chip market with semiconductor giants including Broadcom Inc., Marvell Technologies Inc. and MediaTek Inc., which help companies such as Google LLC, Amazon Web Services Inc., and OpenAI Group PBC design and build their own silicon. The company's ultimate goal is to make chip design more accessible. "Today, state-of-the-art AI reaches most people through a handful of channels and the cost of the compute underneath is the reason for this," Koduri said. "By bringing that cost down, you widen who gets to build with it. Going forward, Koduri said the funds will help the company to finish its first batch of IP and make it available as a product. In addition, it wants to expand its engineering team to try and scale its business. Fundomo partner Rajeev Surati said he's backing Oxmiq because Koduri possesses intimate knowledge of every layer of the silicon stack and knows exactly where the main constraints exist. "Most compute UP makes the customer bend their memory, packaging and foundry around the chip," he said. "Oxmiq does the opposite and that flips a cost center into leverage."

Ex-Intel exec Raja Koduri's OXMIQ raises $35 million to build licensable AI chip architecture

AI chip startup OXMIQ, founded by former Intel executive Raja Koduri, has secured $35 million in Series A funding. The company aims to democratize AI chip development by licensing its GPU architecture, OxCore, allowing businesses and governments to create custom AI processors. This move seeks to reduce the high cost of AI compute, particularly for emerging markets like India and Southeast Asia, fostering wider AI adoption. Hyderabad: AI chip architecture startup OXMIQ, founded by former Intel graphics chief Raja Koduri, has raised $35 million in a Series A funding round to commercialise a licensable graphics processing unit (GPU) architecture that it says will allow companies and governments to build custom AI chips without developing them from scratch. The round, co-led by Samsung Catalyst Fund and Fundomo, takes the company's total funding to $60 million. Existing investors MediaTek, Pegatron Venture Capital, CDIB-TEN, Darwin Ventures, Morgan Creek Digital, AM Intelligence Labs and Intel Capital also participated. Founded in 2024 and headquartered in Campbell, California, with engineering teams in Hyderabad and Bengaluru, OXMIQ is building what Koduri describes as an "Arm for AI GPUs". Just as Arm Holdings licenses CPU designs, OXMIQ aims to license GPU intellectual property, through its flagship product OxCore, enabling customers to build custom AI processors instead of relying solely on Nvidia or investing billions in developing their own silicon. The company is targeting semiconductor firms, neocloud providers, AI infrastructure builders and governments pursuing sovereign AI programmes. The Series A funding will be used to complete OxCore by the end of this year or early next year before customer deployments and commercial silicon tape-outs. Speaking to ET, Koduri said OXMIQ was founded to tackle the rising cost of AI compute, arguing that countries such as India would need to own critical AI chip IP to make AI infrastructure affordable at scale. "India cannot play in AI without taking control of the cost of it. AI compute costs would need to fall by 50-100 times to drive mass adoption," he said adding that chip design is where 92% of margins lie. OXMIQ launched a public beta of its software stack in November 2025 and says it is now being used by 20 companies and 10 universities across nearly 300 GPUs. Koduri said the company's initial commercial focus is India and Southeast Asia, including Malaysia and Indonesia, where AI data centre expansion is driving demand for alternatives to Nvidia-led infrastructure. Earlier this year, OXMIQ partnered with AM Intelligence Labs, part of the AM Green Group, to architect a 2 GW renewable-powered AI compute platform in Uttar Pradesh, with the first 1 GW phase expected to go live by the end of 2027.

OXMIQ raises $35 million to license AI chip designs as sovereign AI demand grows

AI startup Oxmiq has secured $35 million to revolutionize chip design, aiming to slash the costs of building and running AI applications. The company plans to integrate graphics chips, central processors, and tensor engines into a single IP block, aspiring to become the 'Arm' of the AI era. This funding will accelerate the development of their innovative architecture and software, with plans to expand their engineering team. Hyderabad: AI chip architecture startup OXMIQ, founded by former Intel graphics chief Raja Koduri, has raised $35 million in a Series A funding round to commercialise a licensable graphics processing unit (GPU) architecture that it says will allow companies and governments to build custom AI chips without developing them from scratch. The round, co-led by Samsung Catalyst Fund and Fundomo, takes the company's total funding to $60 million. Existing investors MediaTek, Pegatron Venture Capital, CDIB-TEN, Darwin Ventures, Morgan Creek Digital, AM Intelligence Labs and Intel Capital also participated. Founded in 2024 and headquartered in Campbell, California, with engineering teams in Hyderabad and Bengaluru, OXMIQ is building what Koduri describes as an "Arm for AI GPUs". Just as Arm Holdings licenses CPU designs, OXMIQ aims to license GPU intellectual property, through its flagship product OxCore, enabling customers to build custom AI processors instead of relying solely on Nvidia or investing billions in developing their own silicon. The company is targeting semiconductor firms, neocloud providers, AI infrastructure builders and governments pursuing sovereign AI programmes. The Series A funding will be used to complete OxCore by the end of this year or early next year before customer deployments and commercial silicon tape-outs. Speaking to ET, Koduri said OXMIQ was founded to tackle the rising cost of AI compute, arguing that countries such as India would need to own critical AI chip IP to make AI infrastructure affordable at scale. "India cannot play in AI without taking control of the cost of it. AI compute costs would need to fall by 50-100 times to drive mass adoption," he said adding that chip design is where 92% of margins lie. OXMIQ launched a public beta of its software stack in November 2025 and says it is now being used by 20 companies and 10 universities across nearly 300 GPUs. Koduri said the company's initial commercial focus is India and Southeast Asia, including Malaysia and Indonesia, where AI data centre expansion is driving demand for alternatives to Nvidia-led infrastructure. Earlier this year, OXMIQ partnered with AM Intelligence Labs, part of the AM Green Group, to architect a 2 GW renewable-powered AI compute platform in Uttar Pradesh, with the first 1 GW phase expected to go live by the end of 2027.

OXMIQ Raises $35 Million to Scale OxCore™ Architecture

OXMIQ Labs Inc., a unified GPU and AI architecture company founded by Raja Koduri, today closed its $35 million Series A financing, bringing the company's total capital raised to $60 million. The funding will scale OxCore™, OXMIQ's licensable GPU architecture that allows semiconductor companies and AI system builders to build custom AI silicon without a full chip program. 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™), along with the software that runs AI factories and agents. One Core, Three Engines Token demand is outpacing the world's ability to build infrastructure to serve it. OXMIQ was founded to re-architect the GPU stack from Atoms to Agents™, building the silicon IP, configurable systems and software platform that enable semiconductor companies and AI infrastructure builders to drive down the cost of intelligence at every layer of the stack. At the center of the architecture is OxCore™, a scalable, licensable GPU core that integrates three distinct compute engines: a CUDA®-compatible GPU engine, a tensor processing engine, and an orchestration engine (CPU) responsible for coordinating workloads and agents across the system. OxCore tightly couples compute functionality that is typically split across three chips, and was purpose-built for near-memory compute, minimizing data movement to enhance compute and energy efficiency of AI workloads. OxCore was designed for scalability and the architecture scales efficiently from single-core AI deployments to large-scale datacenter configurations. OxCore is running on FPGA today, with live demonstrations available. OxQuilt™, OXMIQ's chiplet integration architecture, combines heterogeneous compute chiplets and memory in a single package. Most AI silicon designs are locked to a specific foundry and memory type. OxQuilt instead adapts to any supply chain, with configuration tools that let customers design across logic process nodes, memory types, interconnect standards, and advanced packaging options. By making high-performance AI compute licensable and configurable, OXMIQ lets any design team build custom AI silicon packages without needing cost-prohibitive full chip programs. The architecture is also designed to incorporate emerging interconnect technologies such as silicon photonics as they reach production readiness. OXMIQ pairs the hardware with a software stack spanning OxCapsule™ for high-level orchestration to low-level kernel optimization. OxPython™ runs existing CUDA® and PyTorch® code on OxCore without code changes, giving developers full portability across hardware. This stack supports emerging silicon architectures for optimized inference at scale and delivers day-zero support for new models. OxPython has been validated on third-party platforms with live demos available. OXMIQ's IP-first model is built for capital efficiency. By focusing on new architecture IP rather than full SoC development, the company generates revenue from customer engagements while preserving capital for building the stack. "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. An Expanding Team OXMIQ has strengthened its board and advisory ranks with two additions that bring decades of silicon pedigree. Jim Keller, CEO of Tenstorrent and among the most influential chip architects in the industry, joins the board of directors alongside existing board member Dr. Ker Zhang. Dr. Valluri (Bob) Rao, a renowned Fellow who retired from Intel's process technology group, joins as an advisor. Together, they deepen OXMIQ's leadership as the company moves from architecture to customer integration. "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. Raja Koduri, OXMIQ founder and CEO, added: "A licensable core with an open architecture means design teams everywhere can build the custom AI silicon their work needs. Today, state-of-the-art AI reaches most people through a handful of channels, and the cost of the compute underneath is the reason. Bring that cost down, and you widen who gets to build with it. I believe AI is a force for good when it is a tool everyone can pick up and use, not just the few who can afford to build with it. Closing this round with investors who own the supply chain tells us we can get there." Get Involved OXMIQ is working with semiconductor companies, neoclouds, AI system builders, and physical AI/ robotics companies ready to own their compute roadmap. For licensing and partnership inquiries, contact [email protected]. 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 multi-stage venture capital fund that invests in deep tech AI infrastructure. MediaTek, a seed investor and one of the world's leading fabless semiconductor companies, is reinvesting. Lawrence Loh, SVP of MediaTek said, "MediaTek is actively powering today's advanced AI capabilities from the edge to the cloud. Our investment in OXMIQ underscores this push and combines our AI ambitions with their highly flexible GPU architecture. We see this investment as a way to continue unlocking unprecedented on-device AI performance across all technology platforms." AM Intelligence Labs, part of the AM Green Group, joins this round as an investor, deepening the collaboration behind a 5GW total AI factory with 3GW as a renewable-powered AI compute platform that they are building with OXMIQ in India. CDIB-TEN is a joint fund between CDIB Capital, a leading Asian PE/VC firm expanding its asset management business through a strong regional presence anchored by over 65 years of investment heritage in Taiwan, and TEN Capital, a prestigious investment fund that is deeply wired into Taiwan's semiconductor ecosystem. Pegatron Venture Capital, the investment arm of one of the world's largest ODMs and system integrators, adds manufacturing and systems depth to carry chiplet AI accelerator designs from architecture to deployment. Morgan Creek Digital, an investor focused on AI and digital infrastructure, joins the round on the thesis that compute capacity and architecture choice will define the AI economy this decade. Darwin Venture Management, a Taipei-based venture firm investing across the TaiwanSilicon Valley technology corridor, brings cross-border conviction that mirrors OXMIQ's own design and supply chain footprint. Intel Capital rounds out the group as a strategic IP partner, adding to OXMIQ's design and engineering depth.

Startup Oxmiq raises $35 million to build chip architecture to lower cost of AI

SAN FRANCISCO, July 1 (Reuters) - Artificial intelligence startup Oxmiq said on Wednesday it raised $35 million from investors in order to build chip design architecture and software that will lower the cost of building and running AI applications. Developing a cutting-edge AI chip can cost hundreds of millions of dollars and take several years to make the requisite silicon design plans and necessary software to support it. Oxmiq's goal is to collapse three distinct components of an AI system and combine them into a single block of IP that it can license, Chief Executive Raja Koduri said in an interview. Typically, AI systems are divided between graphics chips and central processors. Oxmiq plans to combine both, along with a third component, a tensor engine, into a single design, he said. "We would want to be the Arm of this next era," Koduri said, referring to the U.K. company that supplies the design and IP for nearly every smartphone in the world. The company also plans to develop a computing fabric that includes chiplets -- several specific chips combined to form a complete system -- and memory within a single package. The former Intel chief architect and an ex-AMD executive, Koduri said that Oxmiq will also get into the custom chip market, where Broadcom, Marvell and MediaTek compete. Oxmiq plans to use the $35 million to finish the first batch of intellectual property it is working on and make it available as a product, Koduri said. The company also plans to hire more engineers as it scales its business. The Campbell, California, company has raised a total of $60 million, the company said. The $35 million funding round included investors such as Taiwan's MediaTek and Pegatron Venture Capital. Samsung Catalyst Fund and Fudomo led the round. (Reporting by Max A. Cherney in San FranciscoEditing by Nick Zieminski)