All posts

OXMIQ raises $35M to license AI chip designs

Huma ShaziaJuly 1, 2026 at 10:17 PM4 min read
OXMIQ raises $35M to license AI chip designs

Key Takeaways

OXMIQ raises $35M to license AI chip designs
Source: Tech-Economic Times
  • OXMIQ raised $35 million Series A led by Samsung Catalyst Fund and Fundomo, bringing total funding to $60 million
  • The startup is building OxCore, a licensable GPU architecture that lets companies and governments create custom AI processors
  • Initial commercial focus is India and Southeast Asia, with a 2 GW AI compute platform planned in Uttar Pradesh by 2027

OXMIQ, the AI chip architecture startup founded by former Intel graphics chief Raja Koduri, has raised $35 million in Series A funding to build a licensable GPU architecture. The round, co-led by Samsung Catalyst Fund and Fundomo, brings total funding to $60 million. The company plans to complete its flagship product, OxCore, by late 2025 or early 2026.

Koduri's pitch is straightforward: become the ARM of AI GPUs. Just as ARM Holdings licenses CPU designs to chipmakers worldwide, OXMIQ wants to license GPU intellectual property through OxCore. This would let companies and governments build custom AI processors without developing silicon from scratch or depending entirely on NVIDIA.

Advertisement

Who invested and why it matters

The investor list reads like a semiconductor industry reunion. Samsung Catalyst Fund and Fundomo co-led, with participation from MediaTek, Pegatron Venture Capital, CDIB-TEN, Darwin Ventures, Morgan Creek Digital, AM Intelligence Labs, and Intel Capital. That Intel is backing Koduri's new venture after his departure in March 2023 suggests confidence in the business model rather than any hard feelings.

Founded in 2024 and headquartered in Campbell, California, OXMIQ maintains engineering teams in Hyderabad and Bengaluru. The target customers include semiconductor firms, neocloud providers, AI infrastructure builders, and governments pursuing sovereign AI programs.

Why Koduri is betting on emerging markets

Koduri was blunt in his assessment: "India cannot play in AI without taking control of the cost of it. AI compute costs would need to fall by 50 to 100 times to drive mass adoption." He argues that chip design is where 92% of margins lie, which explains why licensing IP rather than manufacturing chips makes sense for countries that want AI infrastructure without NVIDIA-level capital expenditure.

OXMIQ launched a public beta of its software stack in November 2025. The company claims 20 companies and 10 universities are now using it across nearly 300 GPUs. Initial commercial focus is India and Southeast Asia, including Malaysia and Indonesia, where AI data center expansion is creating demand for alternatives to NVIDIA-dominated infrastructure.

2 GW
Planned renewable-powered AI compute platform in Uttar Pradesh, with first 1 GW phase expected by end of 2027
Advertisement

The AM Intelligence Labs partnership

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. The first 1 GW phase is expected to go live by the end of 2027. If completed, this would be one of the largest AI infrastructure projects in India, and a significant proof point for OXMIQ's licensing model.

The project addresses two constraints simultaneously: compute scarcity and energy costs. Renewable power helps manage the economics of large-scale AI workloads, while locally designed chips could reduce dependence on imported hardware.

Can OXMIQ actually compete with NVIDIA?

The honest answer is that OXMIQ is not trying to compete directly. NVIDIA's dominance rests on both hardware and its CUDA software ecosystem, built over 15 years. OXMIQ's bet is that many organizations do not need top-tier NVIDIA performance. They need good-enough AI compute at a price that makes economic sense for their market.

Koduri brings credibility to this bet. He led AMD's Radeon Technologies Group during the Vega and RDNA architecture development, worked on GPU projects at Apple, and ran Intel's Accelerated Computing Systems and Graphics Group, overseeing the Arc GPU initiative. He understands what it takes to build GPU architectures from the ground up.

The risk is execution. Licensing a GPU architecture requires not just working silicon designs but also a software stack that developers actually want to use. NVIDIA's moat is as much CUDA as it is hardware.

ℹ️

Logicity's Take

OXMIQ's licensing model is the interesting part here, not the funding amount. If OxCore delivers even 70% of NVIDIA's inference performance at a fraction of the licensing cost, it opens AI compute to a tier of customers currently priced out. The comparison is not H100 vs. OxCore. It is OxCore vs. no custom silicon at all. For CTOs evaluating AI infrastructure in India or Southeast Asia, the question becomes: wait for OXMIQ to mature, or lock into NVIDIA pricing today? Neither answer is obvious, which makes this worth watching.

Frequently Asked Questions

What is OXMIQ and who founded it?

OXMIQ is an AI chip architecture startup founded in 2024 by Raja Koduri, former Intel graphics chief and AMD Radeon architect. The company is headquartered in California with engineering teams in Hyderabad and Bengaluru.

What is OxCore?

OxCore is OXMIQ's flagship licensable GPU architecture. It allows companies and governments to build custom AI processors without developing chip designs from scratch, similar to how ARM licenses CPU designs.

How much funding has OXMIQ raised?

OXMIQ has raised $60 million total, including a $35 million Series A round co-led by Samsung Catalyst Fund and Fundomo.

When will OxCore be available?

OXMIQ plans to complete OxCore by late 2025 or early 2026, followed by customer deployments and commercial silicon tape-outs.

Which markets is OXMIQ targeting?

OXMIQ is initially focused on India and Southeast Asia, including Malaysia and Indonesia, where AI data center expansion is driving demand for alternatives to NVIDIA-led infrastructure.

Also Read
4,700 lines in 2 days: how AI stopped a WordPress spam flood

Practical example of AI implementation in production environments

ℹ️

Need Help Implementing This?

If you are evaluating AI infrastructure options for your organization, reach out to our team at Logicity for guidance on hardware decisions, cloud vs. on-premise trade-offs, and vendor selection.

Source: Tech-Economic Times / ET

Advertisement
H

Huma Shazia

Senior AI & Tech Writer

Produced with AI assistance and reviewed by the Logicity editorial team. Learn more in our Editorial Policy.

Related Articles