Data is the New Business Currency
summary
Claims that the datasphere (33 ZB in 2018 to 175 ZB by 2025, 80% unstructured) demands a new pipeline: collect → process → store enabled by four things — enterprise GPUs, deep learning, GPU-focused capital, and supporting infra. Describes building GPU-centric data centers using DGX + Pure Storage and high-speed fabric to feed GPUs, running GPU-accelerated Spark. Shows a client cut from 14 GPUs to 2 and sees a 800% speedup, unlocks billion-row analyses, and directly improves KPI visibility.
AI-written summary · about
chapters (29)
- 0:00Intro
- 1:00What this talk is about
- 2:00Data turns into information
- 2:20Who am I
- 3:13Big Data
- 3:55Statistics
- 4:33The Real Story
- 4:54Data Explosion
- 5:13New Paradigm
- 6:14Data Processing Pipeline
- 6:50Gravity
- 7:38Data Pipeline
- 8:20Why Data Pipeline
- 8:37The Human Brain
- 9:16The Academic Aspect
- 9:27Money
- 11:36Why now
- 12:03Four independent technical advances
- 12:21What has changed
- 14:19GPU Architecture
- 15:49What is a GPU
- 16:48AI
- 18:17Deep Learning
- 20:44Supporting Infrastructure
- 21:56Why is it necessary
- 27:27Core Scientific
- 32:41Omniside
- 35:23Fast Data
- 37:01When is it coming
related talks
25mAI-powered digital transformation
38mStop Allocating GPUs, Start Delivering Intelligence: An Enterprise...
33mReimagining the Modern Data Center
12mAnthony Dina, Dell Technologies | Nvidia GTC 2024
20mAI Impact on the Backbone
52mSimplifying the adoption of generative AI for enterprises
25mMaking AI Useful
16mUnleashing the Future of AI Development