Episode summary
Mike Volpi on Why AI Infrastructure Will Outlast the Hype Cycle
The a16z Show
· · Hosts: Raghu Raghuram, Guido Appenzeller · Guests: Mike Volpi
Summary
Mike Volpi argues that the AI buildout rhymes with the late-1990s internet buildout, but with bigger stakes and stronger economics. He says the industry is again pouring capital into foundational infrastructure, yet this time the underlying product is intelligence, not communication, and demand is already visible in the revenue growth of Anthropic, OpenAI, and other model providers.
He expects the market to correct, especially among weaker NeoCloud balance sheets, but thinks the underlying assets will remain valuable for a long time. A key difference from the fiber era is that GPUs appear to be aging more slowly than expected because scarcity, power efficiency, and model demand are extending their useful life. He also notes that AI infrastructure is increasingly bought and financed as a system, not as interchangeable parts.
Volpi says networking will likely have to change as AI workloads become more predictable than internet traffic. He thinks today’s stateless packet-switching assumptions are a poor fit for GPU clusters and that future architectures may look more like circuit-switched, tightly integrated systems. On M&A, he says Cisco succeeded by doing many acquisitions, accepting failure, and using repeatable integration processes; he thinks Nvidia has similar strategic optionality, but that industries still ultimately consolidate on their own terms.
Key takeaways
- Volpi says the AI buildout is likely to be fully useful, but ownership may shift after an economic correction.
- He expects GPUs to have a longer useful life than investors originally assumed because scarcity and token demand keep them productive.
- Volpi says AI networking is misdesigned for stochastic traffic and should move toward more predictable, system-level architectures.
- He argues that Cisco’s M&A success came from doing many deals, accepting failures, and running integration as a repeatable process.
- Volpi says Nvidia’s acquisitions are strategically sensible, but industry structure will still be shaped by scale and consolidation.
Notable quotes
“Do I think the buildout that's happening now is useful and will be fully utilized, if not more, 100%”
“the economic value of intelligence is probably higher than the economic value of communication”
“We all just live in Jensen's world”
“the underlying assets will be quite useful for a long period of time”
“the most fascinating time to be doing our job and to be in our industry”
How the episode opens
We live in a world where demand for AI capacity tokens is near infinite and supply is incredibly constrained. So another way to put it is we all just live in Jensen's world. — You were doing back then M&A for the most valuable company on the planet.
The retro on the M&A process that I think worked well for Cisco was we were not afraid to fail and we did it with high enough frequency that we got a few outliers. You bought this thing for 500 million and like it's doing a billion in earnings 2 years later.
When you see the AI buildout, how do you look back on your previous experiences?
If you look at the revenue ramp of Anthropic, OpenAI are definitely experiencing growth rates that were far superior to what we saw at the time. Do I think the buildout that's happening now is useful and will be fully utilized, if not more, 100%. Will it be owned by the same people that own it today?
Great to have you here on the podcast, Mike. And is a very very opportune time to have you because you were one of the architects of the great internet buildout when you were at Cisco and uh now of course we've got the great AI buildout happening. So uh Vito and I and the audience would love to get your thoughts on then and now uh from all different dimensions from the actual buildout the capitalization of it the technology differences etc etc.
Oh, so let's start. For those that don't uh know or remember, uh Mike was a senior exec at Cisco, overseeing a lot of the M&A and strategic moves that Cisco made in the late 90s. Um and of course that was the time the internet got built out.
From the episode’s captions: roughly the first two minutes, lightly cleaned. Keep watching on YouTube →
Topics
- ai infrastructure
- gpu economics
- data center power
- data center networking
- ma integration
Companies and tickers mentioned
- CSCO
- NVDA
- Cisco
- Nvidia
- Andreessen Horowitz (a16z)
Watch the episode
Watch “Mike Volpi on Why AI Infrastructure Will Outlast the Hype Cycle” on YouTube
Summary, takeaways and quote selection by Octagon. Quotes are short excerpts; the full conversation is in the episode.
More from The a16z Show
-
Inside Salesforce’s Acquisition Strategy for the AI Era -
How AI Is Turning Figma’s Finance Team Into Builders -
The New AI Agents That Shop, Plan, and Gossip -
Andrew Huberman on Focus, Energy, and Performing at Your Best -
What iPhone Packaging Reveals About the Robotics Opportunity -
Hip Hop’s Origin Story, Told by the Legends Who Lived It
The a16z Show, this episode and its artwork belong to Andreessen Horowitz. Octagon is not affiliated with the show; this page is an independent summary.
