Oxmiq Labs has unveiled a new direction in artificial intelligence computing, announcing on Tuesday that it is launching licensable graphics processor technology designed specifically for AI workloads. Founded by Raja Koduri, former chief architect at Intel, the Campbell, California-based startup aims to disrupt the GPU market with a scalable, software-first approach.
The company also confirmed it has raised $20 million in seed funding to support its launch. The round includes investments from angel investors and strategic corporate partners, including semiconductor giant MediaTek. However, Oxmiq Labs did not disclose its valuation.
Unlike traditional chipmakers, Oxmiq is not building physical chips. Instead, it is developing GPU intellectual property (IP) that other companies can license and integrate into their own systems. This model avoids the massive costs of chip fabrication—sometimes exceeding $500 million—while enabling broad adoption across industries.
The technology is highly scalable. It can power a single core for AI applications in robotics and edge devices, or scale to thousands of cores for use in cloud data centers. Furthermore, Oxmiq can customize its GPU architecture to suit specific computing needs, offering flexibility that rivals rigid off-the-shelf solutions.
Koduri, who previously led graphics development at Intel and held senior roles at AMD and Apple, said the company’s vision is ambitious. “We want to be Arm for the next generation,” he told Reuters. Like Arm’s dominance in mobile processor IP, Oxmiq aims to become the go-to architecture for AI-accelerated computing.
A key innovation is Oxmiq’s software-first design strategy. The company has developed a proprietary tool that allows software written for Nvidia’s CUDA platform to run on non-Nvidia hardware—without requiring code changes or recompilation. This could significantly lower the barrier for developers looking to move away from Nvidia’s ecosystem.
Currently, Nvidia dominates the AI GPU market, thanks in large part to CUDA’s entrenched developer base. By enabling seamless compatibility, Oxmiq Labs hopes to attract companies seeking performance without vendor lock-in.
The startup’s focus on IP rather than full chip production allows it to move quickly and adapt to market demands. It also aligns with growing industry interest in modular, licensable technologies—especially as AI drives demand for specialized hardware.
With AI workloads expanding into robotics, autonomous systems, and cloud infrastructure, demand for efficient, scalable GPUs is surging. Oxmiq’s technology could appeal to cloud providers, AI startups, and device manufacturers alike.
While still in its early stages, the company’s leadership, strategic funding, and clear technical vision position it as a potential challenger in the evolving AI hardware landscape.
READ: AI Discovers New Battery Materials for Sustainable Energy
