Jan 2, 2019 · 25m · a16z
a16z Podcast | Datacenter of the Future
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of the a16z Podcast, Chris Dixon and Peter Levine discuss how mobile supply chains, web-scale architecture, and software abstraction are transforming modern data centers and reshaping enterprise IT investment.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
speaking balance: gold is the host, purple is the guest (3 minute bins)
Peter explicitly counters the popular belief that Amazon is leading cloud infrastructure, stating that AWS is built on an older virtual machine generation compared to Google and Facebook.
Hardest push from the host ▶ 12:48 Refusing framing on Amazon's cloud roleChris directly interrupts and pushes back when Peter omits Amazon from leading cloud innovations, arguing that most people consider Amazon the prime leader.
Biggest teaching moment ▶ 13:19 Explaining VM vs container architectural generationsPeter educates Chris on why Amazon's infrastructure is generational tech behind Google/Facebook due to its dependency on virtual machines rather than containerized Linux environments.
The host holds their own ▶ 17:09 Inception analogy of virtual machinesChris demonstrates profound domain knowledge by taking over the conversation to explain the architectural flaws of hypervisors as a 'dream within a dream' hack.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The host as informed peer | Guest teaching | Guest disagreement | The host pushing back | Why |
|---|---|---|---|---|---|---|
| Evolution from Traditional Enterprise IT to Web-Scale Architecture | 7 | 3 | 1 | 2 | Chris Dixon demonstrates strong industry knowledge, outlining historical shifts from Rackspace and Oracle to web-scale architectures like Facebook and Google. He frames the premise that Facebook is more advanced in networking than Cisco, which Peter validates and expands upon. | |
| The Mobile Supply Chain Powering Future Data Centers | 7 | 3 | 1 | 2 | Chris actively synthesizes technical ideas, offering the analogy that a cloud datacenter is essentially reconfigured smartphones using ARM chips. He also explains how startups historically modeled datacenters after Wall Street firms like Fidelity, driving the conversation forward with high expertise. | |
| Software-Defined Networking Portfolio Spotlight: Nicera and Cumulus | 5 | 5 | 2 | 3 | Chris prompts discussions around software-defined networking and storage investments, asking why commodity hardware wasnt adopted earlier. Peter gently corrects/schools him by explaining that x86/PCs were the previous generation's commodity, whereas mobile components mark the current wave. | |
| Future Cloud Models: Private, Public, and Containerized Architecture | 6 | 5 | 2 | 7 | Chris directly challenges Peter's assertion that Google and Facebook are leading over Amazon, pointing out that Amazon is universally recognized as the cloud leader. Peter explains that Amazon relies on older virtual machine architectures rather than modern containerization. | |
| Mesosphere and the Data Center-Scale Operating System | 8 | 2 | 1 | 2 | Chris provides an articulate conceptual breakdown of Mesosphere, comparing datacenter scale OS scheduling to single-server operating system tasks. Peter agrees enthusiastically with the accuracy of Chris's server-to-datacenter analogy. | |
| Virtual Machines versus Containerization | 9 | 2 | 1 | 2 | Chris takes complete command of the topic, explaining containerization versus virtual machines using an Inception-like 'dream within a dream' metaphor. He details deployment friction and how Linux dominance makes server virtualization obsolete. | |
| Hosted Open Source as a Service and New Monetization Models | 6 | 4 | 2 | 5 | Chris questions whether GitHub's business model for open source is a generalizable pattern or an exception due to network effects. Peter argues that turning open source into a managed service unlocks its full market potential. | |
| Podcast Conclusion | 0 | 0 | 0 | 0 | Brief standard podcast sign-off and time limit conclusion. |