The CUDA for Reconfigurable Chips

Advanced, energy-efficient processors like FPGAs are too hard for most developers to program. This hardware-software gap is a major bottleneck, preventing their use in massive markets like edge AI because the required tooling and skills are too specialized.

A unified software platform that acts as a universal compiler. It would provide a simple, high-level abstraction, similar to CUDA for GPUs, letting developers use familiar languages. This makes next-gen hardware accessible and unlocks its true market potential.