Research, product updates, and perspectives on GPU architecture, AI infrastructure, and the future of compute.
Stay in the loop
✓ You're in. Look for the first email shortly.
OXMIQ closes its $35M Series A co-led by Fundomo and Samsung Catalyst Fund, bringing total capital raised to $60 million. Jim Keller joins the board of directors.
Where each product stands, what Q1 brought, and the customer engagement that reframed the licensing-first thesis at scale.
Raja Koduri on energy as the binding constraint in AI compute, sovereign infrastructure, and the path to outcomes-as-a-service.
An MOU to train students across six departments on GPU architecture, AI systems, and the silicon that powers modern compute.
Strategic partnership to develop 2 GW AI compute infrastructure by 2030, beginning with a 1 GW facility in Uttar Pradesh, India.
In conversation with Raja Koduri on why the GPU stack needs to be rebuilt from the ground up for the agentic era.
Exploring a machine with four RTX 6000 Blackwell GPUs delivering 384 GB of VRAM for local LLM inference workloads.
Memory bandwidth and economics, not just compute, determine the future of AI architecture. A look at what comes next.
How chiplet-based design enables scalable, composable silicon for the inference era and why it changes GPU economics.
OxCapsule is now in public beta. Authenticate once and connect to GPU compute anywhere in under 15 seconds, across VS Code, Cursor, SSH, and Docker with native GPU access.
OXMIQ's launch and Raja Koduri's vision for India's growing role in the global AI and semiconductor landscape.
Raja Koduri lays out OXMIQ's vision for a coherent, open GPU stack, from silicon IP all the way to the developer CLI.