A startup says the AI bottleneck isn’t compute. It’s memory, and it ditched the GPU to prove it.
SMRTR summary
Majestic Labs, a Tel Aviv startup, claims the real AI bottleneck is memory, not computing power. Its Prometheus server ditches Nvidia GPUs entirely, using custom processors paired with up to 128TB of affordable phone-grade memory, promising 50 times more fast memory than Nvidia's top system. Hardware hasn't shipped yet.
SMRTR provides this summary for quick context. The original article belongs to The Next Web.
Read the original article