Most AI infrastructure today is hitting a breaking point. Marc Austin, CEO of Hedgehog, reveals how open source networking and cloud-native solutions are revolutionizing how enterprises build and operate AI at scale. This episode addresses issues many building AI infrastructure today are facing — expensive proprietary systems, overwhelming complex network configurations, and ways to make on-prem AI infrastructure feel just like the public cloud.
We discuss how networking is the hidden bottleneck in scaling GPU clusters and the surprising physics and hardware innovations enabling higher throughput. Marc shares the journey of building Hedgehog, an open source, cloud-native platform designed for AI workloads that bridges the gap between complex hardware and seamless, user-friendly cloud experiences. Marc explains how Hedgehog's software abstracts and automates the networking complexity, making AI infrastructure accessible to enterprises without dedicated networking teams.
We break down the future of AI networks, from multi-cloud and hybrid environments to the rise of Neo Clouds and the open source movement transforming enterprise AI infrastructure. If you're a CTO, data scientist, or AI innovator, understanding these network innovations can be your moat. Listen to this episode to see how open source, cloud-native networking, and physical innovation are shaping the AI infrastructure of tomorrow.
Podcast Links
Watch: https://www.youtube.com/@alexa_griffith
Read: https://alexasinput.substack.com/
Listen: https://creators.spotify.com/pod/profile/alexagriffith/
More: https://linktr.ee/alexagriffith
Website: https://alexagriffith.com/
LinkedIn: https://www.linkedin.com/in/alexa-griffith/
Find out more about the guest at
LinkedIn: https://www.linkedin.com/in/austinmarc/
Website: https://hedgehog.cloud/
Github: https://github.com/githedgehog
Chapters
00:00 Rethinking AI Infrastructure
02:49 The Role of Networking in AI
05:54 Marc's Journey to Hedgehog
08:46 Lessons from Big Companies
11:38 Requirements for AI Networks
14:48 Advancements in AI Networking
17:33 Future Challenges in AI Infrastructure
20:46 Creating a Cloud Experience On-Prem
23:32 The Shift to Hybrid Multi-Cloud
28:10 Evolving AI Infrastructure and Efficiency
30:57 AI Workloads and Network Configurations
32:41 Zero Touch Lifecycle Management
35:12 Support for Hardware Devices
35:45 Networking Paradigms and Vendor Lock-in
38:42 The Rise of Neo Clouds
41:31 Demand for AI Infrastructure
43:57 Open Source and Cloud-Native Networking
47:27 Challenges of Building a Networking Startup
50:46 Proud Accomplishments at Hedgehog
52:41 Future Excitement in AI Inference
