The AI chip war just hit warp speed.

AMD’s Instinct MI355X has officially crushed the MLPerf Inference 6.0 benchmarks, obliterating the one million tokens per second threshold.

This isn't just a lab win—it’s a massive leap for enterprise-grade LLM performance.

With support from giants like Dell and HPE, AMD is officially challenging NVIDIA’s throne in the high-stakes AI accelerator arena.