WeAreDevelopers Logo WeAreDevelopers
Sign in
For You LIVE
World Congress 2024 • Aug 20, 2024 • Session details

AI Factories at Scale

Thomas Schmidt

Generative AI demands a shift from legacy CPUs to high-density GPU clusters. Learn how to architect scalable AI factories that slash compute energy usage.

Pause
Mute Enter Fullscreen
#1 about 3 min

Amber's evolution as a pioneer in GPU acceleration

How early fluid dynamics research evolved into a foundational enterprise partnership with Nvidia for AI infrastructure.

#2 about 3 min

Milestones driving the rapid evolution of generative AI

Landmark computing breakthroughs like CUDA and transformers dramatically accelerated the foundational capabilities of artificial intelligence models.

#3 about 3 min

Financial impacts of generative AI across the enterprise

Pushing early generative artificial intelligence experimentation into real-world use cases creates significant return on investment for enterprises.

#4 about 2 min

Contrasting legacy supercomputers with modern AI clustered infrastructure

The modern DGX H100 provides exponential performance improvements at a fraction of the cost and compute footprint of traditional supercomputers.

#5 about 3 min

Replacing legacy CPUs to reduce data center energy consumption

Transitioning high-intensity workloads from CPUs to specialized GPUs significantly lowers global energy constraints while preventing thermal overload.

#6 about 2 min

Escalating compute demands for production generative AI inference

The transition toward production software integration multiplies infrastructural demands required to serve transformer models asynchronously.

#7 about 3 min

Core infrastructure components required for an AI factory

Essential deployment layers combine internal cluster networking, advanced storage structures, and modular temperature controls to finalize AI architecture.

#8 about 1 min

Deploying an immediate AI center of excellence via superpods

Turnkey clusters establish scalable reference structures directly optimized for massively parallel application training deployments.

#9 about 5 min

Managing the AI cluster using stack synchronization software

Dedicated management software resolves complex stack synchronization, coordinates hybrid cloud workloads, and proactively monitors cluster networking health.

#10 about 1 min

Allocating GPU resources and implementing multi-tenant usage chargebacks

Defining specific multi-instance constraints permits scalable resource allocation while allowing enterprises to accurately attribute processing costs to specific users.

#11 about 2 min

High-performance storage necessities for continuous large LLM training

Specialized high-bandwidth storage tiers bypass input bottlenecks that frequently stall training sequences during comprehensive model checkpointing.

#12 about 3 min

Adapting data center environments for direct liquid immersion cooling

Upgrading physical infrastructure arrays toward integrated immersion tanks safely manages the vast thermal footprints generated by emerging inference parameters.

Matching moments

1:25 min

Addressing the sustainability and power consumption of AI

Christian Heilmann Christian Heilmann · World Congress 2026 Europe

2:12 min

Transitioning from data centers to AI factories

Michael Kagan Michael Kagan +1 · World Congress 2026 Europe

1:29 min

Tech infrastructure capacity and AI product innovations

4:03 min

Managing massive power consumption scaling in AI data centers

Stephan Gillich Stephan Gillich +3 · World Congress 2024

2:35 min

Balancing AI competitiveness with compute efficiency demands

Markus Hacker Markus Hacker +1 · World Congress 2025

3:57 min

Strategies for accelerating innovation and maximizing AI value

Stephan Gillich Stephan Gillich · World Congress 2024

More from World Congress 2024

Beyond the Hype: Real-World AI Strategies Panel thumbnail
24:07
12 months ago

Beyond the Hype: Real-World AI Strategies Panel

Mike Butcher, Jürgen Müller, Katrin Lehmann, Tobias Regenfuss

Carl Lapierre - Exploring Advanced Patterns in Retrieval-Augmented Generation thumbnail
21:17
almost 2 years ago

Carl Lapierre - Exploring Advanced Patterns in Retrieval-Augmented Generation

Carl Lapierre

Coding an Immersive Copilot using Unity / .NET and Azure OpenAI! thumbnail
31:50
about 2 years ago

Coding an Immersive Copilot using Unity / .NET and Azure OpenAI!

Zaid Zaim

Building a diverse tech company: The power of Diversity, Equity, Inclusion, and Belonging thumbnail
29:43
about 2 years ago

Building a diverse tech company: The power of Diversity, Equity, Inclusion, and Belonging

Sue Quackenbush

Dapr in Practice thumbnail
24:08
about 2 years ago

Dapr in Practice

marcklefter

Minimising the Carbon Footprint of Workloads thumbnail
27:46
about 2 years ago

Minimising the Carbon Footprint of Workloads

Michael Mueller

Related videos

Your Next AI Needs 10,000 GPUs. Now What? thumbnail
29:52
World Congress 2025 • about 1 year ago

Your Next AI Needs 10,000 GPUs. Now What?

Anshul Jindal, Martin Piercy

Building the Nervous System of AI - Michael Kagan (NVIDIA) thumbnail
29:17
World Congress 2026 Europe • 3 months ago

Building the Nervous System of AI - Michael Kagan (NVIDIA)

Michael Kagan, Sead Ahmetović

Bringing AI Everywhere thumbnail
26:25
World Congress 2024 • about 2 years ago

Bringing AI Everywhere

Stephan Gillich

The Future of Computing: AI Technologies in the Exascale Era thumbnail
30:46
World Congress 2024 • about 2 years ago

The Future of Computing: AI Technologies in the Exascale Era

Stephan Gillich, Tomislav Tipurić, Christian Wiebus, Alan Southall

Efficient deployment and inference of GPU-accelerated LLMs​ thumbnail
23:01
World Congress 2024 • about 2 years ago

Efficient deployment and inference of GPU-accelerated LLMs​

Adolf Hohl

Beyond the Hype: Real-World AI Strategies Panel thumbnail
24:07
World Congress 2024 • 12 months ago

Beyond the Hype: Real-World AI Strategies Panel

Mike Butcher, Jürgen Müller, Katrin Lehmann, Tobias Regenfuss

Related articles

From learning to earning

Jobs that call for the skills explored in this talk.