{"data":{"slug":"nvidia-b300","etype":"chip","name":"NVIDIA B300 (Blackwell Ultra)","cat":"GPU","subtitle":"Blackwell Ultra: 288 GB of HBM3E and ~1.5x the FP4 for the reasoning-inference era.","summary":"Announced at GTC 2025 as Blackwell Ultra, the B300 raises memory to 288 GB using 12-high HBM3E stacks and lifts dense FP4 throughput roughly 1.5x over B200 — aimed squarely at long-context, high-token reasoning inference. It ships in GB300 NVL72 racks from late 2025 into 2026.","attrs":[{"g":"Overview","l":"Designer","v":"NVIDIA"},{"g":"Overview","l":"Architecture","v":"Blackwell Ultra"},{"g":"Overview","l":"Announced","v":"March 2025 (GTC); systems from late 2025"},{"g":"Overview","l":"Status","v":"Ramping"},{"b":"Vendor spec","g":"Memory","l":"Memory","v":"288 GB HBM3E (12-high stacks)"},{"b":"Vendor claim","g":"Compute","l":"FP4 throughput","v":"~1.5x B200 (vendor figure)"},{"b":"Reported","g":"Power","l":"TDP","v":"up to ~1,400 W in GB300 configurations"},{"b":"Vendor spec","g":"Physical","l":"Forms","v":"GB300 NVL72 rack systems; HGX B300"}],"kf":{"node":"TSMC 4NP","year":2025,"class":"GPU","tdp_w":1400,"mem_bw":8.0,"memory":"288 GB HBM3E","designer":"NVIDIA"},"aliases":["B300","Blackwell Ultra","GB300"],"ticker":null,"exchange":null,"hq":null,"founded":null,"website":null,"status":"Ramping","region":null,"asof":"2026-01-15","sources":[{"url":"https://www.nvidia.com/en-us/data-center/gb300-nvl72/","label":"NVIDIA GB300 NVL72 page"}]},"meta":{"found":true,"count":1,"served_from":"site cache (no live upstream calls)","as_of":"2026-10-14T11:46:09+00:00","disclaimer":"Research and information only — not investment advice. Structured records carry basis labels (vendor spec / reported / estimate); market quotes are delayed/indicative.","attribution":"Records compiled from vendor documentation, official filings and reputable public reporting; market quotes via financial data providers. See /data-sources."}}