---
id: "entity-nvidia-blackwell"
type: "entity"
source_timestamps: ["§ Transforming Moats"]
tags: ["hardware", "compute"]
related: ["claim-compute-scaling-rate"]
entityType: "product"
canonicalName: "NVIDIA Blackwell"
aliases: ["Blackwell", "Blackwell GPU", "NVIDIA Blackwell GPU"]
sources: ["futures"]
sourceVaultSlug: "hbr-seg-futures"
originDay: 2
articleStem: "hbr-nm-99-genai-end-incumbent-advantage"
sourceUrl: "https://hbr.org/2024/11/could-gen-ai-end-incumbent-firms-competitive-advantage"
sourceTitle: "Could Gen AI End Incumbent Firms’ Competitive Advantage?"
---
# NVIDIA Blackwell GPU

**Profile.** A state-of-the-art GPU architecture by NVIDIA (successor to Hopper), designed for AI and accelerated computing and marketed for extreme training and inference workloads.

**Role in the source.** Used to illustrate the exponential growth of compute power in [[claim-compute-scaling-rate|the 4× Moore's Law claim]]: the seminal **2012 AlexNet** model (which took *a week* to train on *two GPUs*) could be trained on a **single Blackwell GPU in about 5 minutes**.

**Canonical reference:** NVIDIA product page. *(Enrichment note: modern flagship GPUs are orders of magnitude more powerful than the GTX 580s used for AlexNet; the specific 5-minute figure is illustrative rather than documented in the enrichment set.)*