Episode Summary
Executive Summary: Martin Casado argues AI is not in a demand bubble but in a supply-constrained expansion: real usage, budgets, and productivity gains are surging faster than compute, data centers, power, and regulation can keep up. He says coding is commoditizing, engineering is not, SaaS will be reshaped mainly at the consumption layer, and the biggest enterprise disruption may be invisible decision-making by agents.
Main Topics: AI demand vs. supply constraints (Priority: 5/5): The conversation rejects the idea of an AI demand bubble and instead frames the market as under-supplied. Demand is real, monetization is happening, and bottlenecks increasingly sit in infrastructure, permitting, and power rather than model quality. Infrastructure is back at the center (Priority: 5/5): Casado argues every technical epoch forces a rebuild of the stack. AI is reviving interest in chips, networking, storage, and data-center infrastructure, making infrastructure again a core strategic layer. AI and the future of coding/engineering (Priority: 5/5): He distinguishes between 'coding' becoming dramatically easier and engineering remaining essential. AI lowers the floor for developers but does not lower the ceiling, especially for large, stateful, operational systems. Enterprise software and SaaS durability (Priority: 5/5): The speakers debate whether AI kills SaaS. Casado says SaaS is not fundamentally hard to build; it persists because it encodes business processes, compliance, and operational reality. AI changes how users interact with software, not the need for structured systems. Decision-making shifts from humans to agents (Priority: 4/5): A major unresolved issue is who makes technical purchasing and infrastructure decisions when agents are writing code and selecting tools. This could alter enterprise buying, central procurement, platform teams, and IT governance. Pricing and monetization shift to usage (Priority: 4/5): The panel suggests software pricing may move from seat-based or recurring licensing toward consumption models based on tokens and actions, echoing the disruption from perpetual licenses to SaaS subscriptions. Regulation as the main bottleneck (Priority: 5/5): Casado repeatedly insists the real constraint on scaling AI infrastructure is regulatory delay—especially permitting and breaking ground for data centers—not technical feasibility. He contrasts U.S. bureaucracy with China’s speed.
Key Arguments: AI demand is real: companies are deploying models, budgets are shifting, and productivity gains are already visible. The current issue is a supply underhang, not a supply overhang; compute, power, and data-center capacity lag demand. Every major technical epoch requires rebuilding much of the stack, which is why chips and networking are newly important again. AI makes coding easier for more people, but engineering remains necessary because complex, stateful systems and operations are still hard. The biggest AI-driven disruption to enterprise software is likely at the consumption layer, not the business-process layer. SaaS endures because buyers are purchasing business workflows, compliance, and operational guarantees—not just UI software. Agents selecting tools may reduce the visibility of the human decision-maker, creating new questions for central buyers, IT, and platform teams. The largest constraint on infrastructure scaling is regulatory delay in permitting/building, not lack of technical capability or capital. Software pricing is likely to migrate toward token/action-based consumption models as agentic usage grows. Public-market weakness in legacy software reflects budget movement to new AI-related areas, not a simple collapse in software value.
Data Points: Podcast/episode reference: 6-5 podcast / A16Z feed drop - The transcript introduces this as a feed drop from the 6-5 podcast featuring Martin Casado. A16Z tenure: 10 years - Casado says he has been at Andreessen Horowitz for 10 years. Tech experience: 30 years - Casado references having spent 30 years in tech while discussing bubbles and market rationality. Public segment count: 50th TV segment - Daniel Newman says he has done his 50th TV segment that week answering questions about an AI bubble. Pricing model shift: from perpetual license to recurring, then recurring to consumption - The panel compares historical software pricing transitions to the coming AI-era move toward consumption-based pricing. Time horizon: a couple of years in - Casado says AI is early but sufficiently established to draw some speculative conclusions. Infrastructure scale: multi-trillion dollar business - Casado describes infrastructure as a multi-trillion-dollar business when discussing agentic decision-making. Regulatory example fine: $17 million - The hosts cite Italy fining Cloudflare as an example of regulatory and compliance pressure on tech companies.
Pivotal Quotes: "It's very clear that coding is pretty much dead, but engineering is very much not." — Martin Casado: He distinguishes between the ease of generating code and the continuing need for systems engineering, operations, and complex software design. "We do not have a supply overhang, we have a supply underhang." — Martin Casado: Casado rejects the AI bubble framing and says demand is outpacing infrastructure supply. "The biggest discussion there last year was we over-regulated ourselves." — Daniel Newman: He is referring to Davos and the broader policy discussion about Europe slowing technology companies through regulation.
Implications: AI is likely to expand, not collapse, but its growth will be limited by infrastructure, power, and permitting. Enterprises should expect changing buying behavior, agentic procurement, and a shift toward usage-based software models.
From the Transcript
What's going to happen to central buyers and platform teams and IT teams if agents are making the decision? It's very clear that coding is pretty much dead, but engineering is very much not. Every time you have a technical epoch, you have to redo everything, and we forget that every time. I don't think people even have a common definition of a bubble. If AI demand is real and Excel, Why does everything still feel constrained? Why does a technology that's clearly delivering value also feel harder to scale than expected? We've seen this pattern before. In early technology shifts, it was easy to assume the hard problems were solved. Infrastructure was treated as finished, then usage surged. Systems built for a smaller world began to fail. Networks drained, power, physical footprint, and coordination became first-order constraints again. Each new technical epoch forced a rebuilding of the stack. AI is creating that moment now. The demand is not speculative. Companies are deploying models, budgets are moving. Real productivity gains are already showing up, and yet nearly every part of the system feels tight. Compute is scarce, data centers take years to permit and build, power is difficult to secure. Regulation moves far more slowly than the technology itself. This has led to two dominant stories. One says we're in an AI bubble. The other assumes scale will smooth everything out. Neither fully explains.
Demand is real. You have real users paying real money, getting real value, and that's incredibly clear. The data couldn't be more clear. So, is there like a demand bubble, meaning like we're assuming a demand will come? Do we have a supply overhang where we're building out as hoping it'll come? No, the answer is absolutely not. We do not have a supply overhang, we have a supply underhang. The demand is very real. Now, speculatively, if you look on a deal-by-deal basis, Sure, some deals are overvalued, but some deals are undervalued. So, what I've learned in 10 years of investing, 30 years in tech, is that markets are actually very rational in the long term and broadly, but it's uneven. So, if depending on how you look at it and how you squint, you're going to see things that seem overvalued and things seem undervalued. But I would say, if you take it all in, my true belief is it's all undervalued in the long term. This stuff is so, so disruptive. Demand is.
A lot of your compatriots are actually. I'll be going the other way. No, no, I hear you. It's funny. A lot of your peers are headed there for the first time in years. And Daniel and I, for the first, you know, we just started going last year when it looked like we were actually going to talk about technology. And yeah. So, yeah, it's big word to them, brother. I was thinking about your regulation, your regulation commentary, but quite frankly, the biggest discussion there last year. Was we over-regulated ourselves, and these were uh people from the uh from the EU and DJT is going to be there and uh Vance and folks. I mean, did you just see like it like Italy just fined Cloudflare, what, 17 million dollars for like something that they basically couldn't fix? And Matt Prince is gonna find. I mean, I just feel like we're in this crazy thing where we're extracting, like we're slowing down companies taxation. This is what this is.
About The a16z Podcast
The a16z Podcast discusses tech and culture trends, news, and the future – especially as ‘software eats the world’. It features industry experts, business leaders, and other interesting thinkers and voices from around the world. This podcast is produced by Andreessen Horowitz (aka “a16z”), a Silicon Valley-based venture capital firm. Multiple episodes are released every week; visit a16z.com for more details and to sign up for our newsletters and other content as well!