The latest generation of AI models achieve breakthrough performance by using vastly more computation to solve complex problems. As it stands, the world will not be able to meet demand for compute as frontier models become ever more sophisticated.

Even the most dominant players, with full-stack control across silicon, software, and supply chains, are unable to solve this within the existing architecture.  Inherent constraints in physical design and packaging mean GPU memory architecture is incapable of simultaneously delivering high throughput, high interactivity and low cost.

Continuing AI’s advance requires a new compute paradigm. One that can overcome the fundamental limits of memory, energy, and speed that define today’s systems.