The commercial home of
Surface AI in silicon.
SurfaceEDGE. Custom AI semiconductors. IoT devices. Core licensing. Advanced AI R&D and manufacturing partnerships — from Austin.
Today's LLM stack was built for hyperscale GPU-cloud deployment.
AI inference requires huge amounts of high speed memory to hold the answer while the AI works it out and vast compute power to manage it all quickly. Current AI is too heavy, too power-hungry, too memory-intensive and too cloud-dependent.
SurfaceEDGE changes that.
SurfaceEDGE AI designed for silicon.
SurfaceEDGE is built for the opposite world. Same inference but in a tiny memory footprint so it needs far less processor power — resulting in full AI running on low-power devices, industrial controllers, autonomous systems, sovereign field deployments, vehicles, robotics, secure facilities and AI-on-Chip.
If SurfaceCORE turns standard enterprise compute into AI infrastructure, SurfaceEDGE goes further — and with the quad core AI processor it's even faster than SurfaceCORE. It pushes AI into the vast embedded, edge, industrial, sovereign and silicon markets that current LLM architectures cannot economically or physically access.
That is the real opportunity.
Surface inside the device.
The commercial edge AI product. Industrial PCs, IoT devices, and custom silicon. One device. One Geode. No cloud.
→3D IC. Quad-node. Purpose-built for Surface.
Tri-core AI processor + DSP co-processor. 700MB mask ROM foundation weight. 1GB non-volatile learning core. Silicon accelerated Surface AI for true embedded standalone AI applications or accelerated Surface AI blades for data center deployments.
→Surface in your product.
Core licensing for OEMs, platform vendors, and system integrators embedding SurfaceEDGE in their own hardware.
→The next generation of AI silicon.
Advanced AI semiconductor research — 3D IC architecture, on-chip learning, energy-efficient edge inference.
→Large language models will
never run at the edge.
Surface already does.
Too large. Too power-hungry. Too cloud-dependent. The AI silicon market at the edge — defence, industrial, autonomous, IoT at scale — is currently without a production-ready deterministic AI silicon solution.
1,675 Surface nodes run on the energy of one large AI model. That's the power budget that edge and autonomous systems actually have.
Ready to build with
Surface in silicon?
Partner with us →Licensing, silicon design partnerships, and investment conversations open.