
Frontier-class AI. Right on your desk. What happens when you cluster four @Lenovo ThinkStation PGX systems? You get the power to run a 552B-parameter AI model without quantization at 70–200 tokens per second — no external QSFP switch required. Plus, see how a PGX cluster can bring the cloud on-premises with a live AI tokenomics demo with @NVIDIAworkstatn at #SC26. Big AI performance. Small footprint. Huge possibilities.







