Majestic Labs Takes Aim at Nvidia with Arm-Powered AI Server

Majestic Labs, an Israeli startup founded in 2023 by former Google and Meta engineers, has created a server that directly competes with Nvidia's GPU offerings. The company argues that pairing expensive graphics processors with high-bandwidth memory is no longer the best approach for AI inference.

The Prometheus server replaces GPUs with Ignite AI Processing Units (AIUs), which combine Arm cores with RISC-V vector and tensor engines. Each Prometheus server can house up to 12 AIUs, sharing between 8 TB and 128 TB of LPDDR6 memory across one contiguous, coherent pool.

This architecture allows for a significant increase in memory capacity, with Majestic claiming it delivers over 50 times more fast memory than Nvidia's DGX B300 system. In fact, the company says that one Majestic rack can hold the equivalent of 25 Nvidia NVL72 Vera Rubin racks at a fraction of the power.

Majestic Labs estimates that the Prometheus server could cost between 10 and 50 times less than a GPU system of equivalent performance once it ships next year. The server is designed to be OCP-compliant and will support PyTorch, vLLM, and OpenAI's Triton frameworks.