Why We Don't Need More Data Centers - Sustainable AI Infrastructure Through GPU Marketplaces
AI Engineer via YouTube
Learn Python with Generative AI - Self Paced Online
AI, Data Science & Cloud Certificates from Google, IBM & Meta
Overview
Google, IBM & Meta Certificates — All 10,000+ Courses at 40% Off
One annual plan covers every course and certificate on Coursera. 40% off for a limited time.
Get Full Access
Explore why the current approach of continuously building more data centers and deploying additional GPUs represents a fundamentally flawed, expensive, and environmentally unsustainable strategy for AI infrastructure in this 14-minute conference talk. Learn how the endless cycle of data center expansion masks deeper infrastructure inefficiencies and discover why leveraging GPU marketplaces for dynamic resource allocation offers a more sustainable solution. Examine practical use cases demonstrating how companies can scale GPU capacity flexibly, how startups can access affordable compute resources, and how idle GPUs can be monetized effectively. Understand the vision for a future of sustainable and democratized AI infrastructure that moves beyond traditional data center expansion. The presentation, delivered by Dr. Jasper Zhang, CEO and Co-founder of Hyperbolic, draws on his expertise as a mathematical prodigy with a Ph.D. from UC Berkeley and experience at Ava Labs and Citadel Securities, recorded at the AI Engineer World's Fair in San Francisco.
Syllabus
Why We Don’t Need More Data Centers - Dr. Jasper Zhang, Hyperbolic
Taught by
AI Engineer