The press release from OXMIQ Labs announces a $35M Series A round co-led by Samsung Catalyst Fund and Fundomo to scale...
The press release from OXMIQ Labs announces a $35M Series A round co-led by Samsung Catalyst Fund and Fundomo to scale OxCore, a licensable GPU architecture that includes a CUDA-compatible engine.
confidence score
Strong evidence: 4 independent source classes support this read.
signal brief
The press release from OXMIQ Labs announces a $35M Series A round co-led by Samsung Catalyst Fund and Fundomo to scale OxCore, a licensable GPU architecture that includes a CUDA-compatible engine. OxPython allows existing CUDA and PyTorch code to run without modification, directly threatening Nvidia's CUDA monopoly. The funding total reaches $60M, with participation from MediaTek and others. Manifold Markets consensus shows 56.95% probability that CUDA remains a monopoly through 2027, indicating market awareness of competitive risks. While early-stage, the architecture is validated on FPGA and backed by a team led by Raja Koduri, formerly at Intel and AMD. This signal points to potential erosion of CUDA's lock-in over the medium term.
What the sources said:
- OXMIQ press release: "OxPython runs existing CUDA and PyTorch code on OxCore without code changes, giving developers full portability across hardware."
- Manifold Markets: Consensus at 56.95% for "Will CUDA remain a monopoly for GPU software through 2027?"
- HN Show HN: "NanoEuler – GPT-2 scale model in pure C/CUDA from scratch" – indicates ongoing developer engagement with CUDA but also alternative implementations.
source data used
“Score: 2 | Answers: 0 | Views: 41 Tags: python, pytorch, cuda CUDA error: device-side assert triggered" during backward pass, but error points to an unrelated .to(device) call”
“Score: 2 | Answers: 0 | Views: 60 Tags: python, pytorch, parallel-processing, cuda, cuda-streams How can I leverage torch.cuda.Stream() to control CUDA streams for parallel task execution?”
“Points: 55 | Comments: 26 Author: vforno Link: https://github.com/JustVugg/nanoeuler Show HN: NanoEuler – GPT-2 scale model in pure C/CUDA from scratch”
“Manifold consensus on 'Will CUDA remain a monopoly for GPU software through 2027?': YES=56.95%”
“Off the Wire Press Releases CAMPBELL, Calif., July 2, 2026 — OXMIQ Labs Inc., a unified GPU and AI architecture company founded by Raja Koduri, has closed its $35 million Series A financing, bringing the company’s...”
Decision support, not stock advice. This signal is research with cited evidence — not a recommendation to buy, sell, or hold any security.