OXMIQ's pitch, a single licensable GPU core spanning graphics, tensor, and orchestration compute, matters to hardware and ML infrastructure practitioners because it targets the same software-compatibility bottleneck that keeps most non-Nvidia silicon locked out of the CUDA/PyTorch ecosystem. Per Reuters, OXMIQ Labs, founded by former Intel chief GPU architect Raja Koduri, closed a $35 million Series A (total raised $60 million), co-led by Samsung Catalyst Fund and Fundomo, with MediaTek and Pegatron Venture Capital participating. Koduri told Reuters OXMIQ wants to "be the Arm of this next era" of chip design. The company's OxCore IP uses RISC-V cores and near-memory compute concepts and is paired with an OxPython software layer designed to run existing CUDA and PyTorch code without modification, a bid to solve the software-lock-in problem that has hobbled prior Nvidia challengers.
Startup Oxmiq raises $35 million to build chip architecture to lower cost of AI