AMD claims MI350P delivers 2–5x tokens per dollar in enterprise workloads
ryanshrout · x · 2026-07-24
AMD said the MI350P delivers 2–5x tokens/sec/$ across enterprise workloads.
On stage, AT&T was highlighted as a deployment example, with AMD showing the chip in the context of large-scale enterprise AI adoption. The slide deck also repeated the pitch that enterprise buyers care about power, cost, and ease of deployment.
More from Companies & People
- David Sacks Calls to End 'Kimi Panic', Asserts US Frontier Models Still Lead — Dan_Jeffries1 · 2026-07-24
- AI Market Dynamics: Capital Backing OpenAI and Anthropic vs. The Rest — tekbog · 2026-07-24
- NeurIPS re-releases OpenReview reviews after a technical issue, taking about 4 hours — NeurIPSConf · 2026-07-24
- VC firms are creating a Chief AI Officer role — gamersecret2 · 2026-07-24
- Tolan CEO explains why voice, memory and even bad AI poetry shaped the companion app — round · 2026-07-24
- Runway schedules an NYC industry night with pizza, drinks, and networking — runwayml · 2026-07-24