This episode features Martine Casado, General Partner at Andreessen Horowitz, where he runs the infrastructure fund focusing on compute, networking, storage, databases, AI infrastructure, dev tools, and security. Prior to A16Z, Martine was a portfolio founder and has been investing in infrastructure and enterprise for over 10 years.

The conversation explores the current AI infrastructure landscape, examining whether we're experiencing an AI bubble or supply constraints. Martine argues that demand for AI is real and measurable, with companies paying for genuine productivity gains, but supply constraints - particularly regulatory bottlenecks - are the primary limiting factors.

Key topics include the transformation of software development (coding vs. engineering), the future of SaaS business models, enterprise software disruption, and infrastructure scaling challenges. The discussion reveals how AI agents are beginning to make technical decisions traditionally handled by IT teams, potentially reshaping trillion-dollar infrastructure markets.

AI Demand Reality vs. Bubble Speculation

"From a productivity standpoint, demand is real. You have real users paying real money, getting real value, and that's incredibly clear" - Martine argues against AI bubble concerns

"We do not have a supply overhang. We have a supply underhang" - demand consistently outpaces supply across the AI infrastructure stack

"Markets are actually very rational in the long term and broadly, but it's uneven" - some deals overvalued, others undervalued, but overall undervalued long-term

The Death of Coding, Not Engineering

"It's very clear that coding is pretty much dead, but engineering is very much not" - AI lowers the floor for development but doesn't lower the ceiling

"The companies that are the most aggressively using AI are also hiring the most" - contradicting expectations that AI would reduce engineering headcount

AI struggles with "large, complex, stateful software code bases" and operations - running software in production remains unsolved

"The tent gets a lot bigger" - more people can code, requiring more operations and professional developers for harder problems

SaaS Disruption: Business Process vs. Technology

"SaaS has never been a technology problem ever. It's just not hard" - the value is in encoding business processes, compliance, and operational reality

"Why do people buy SaaS? And the answer is you're buying a business process" - not the underlying software technology

AI changes consumption layers (natural language interfaces) but doesn't eliminate the need for structured data, compliance, and complex business processes

"We're seeing the largest move from recurring to consumption basis" pricing - comparable to the perpetual license to SaaS transition

Agent Decision-Making Disrupts IT Infrastructure

"What's going to happen to central buyers and platform teams and IT teams if agents are making the decisions?" - fundamental shift in infrastructure purchasing

"Infrastructure is a multi-trillion dollar business, and you've removed the human by and large from actually making the decision of what to use"

Traditional model: developers request databases from IT teams who provide approved options with organizational policies

"The AI is making that decision" when developers use tools like Cursor or Claude for coding - bypassing human infrastructure choices

Regulatory Bottlenecks Trump Technical Constraints

"There's only one constraint... It is so onerous to break ground in the United States. It makes more sense to go to space"

"Data centers in space" concept pencils out financially "just because of regulation" - highlighting regulatory burden severity

"The long pole by far, by order of magnitude is breaking ground. That's it. We know how to solve power. We know how to build foundries"

"Is China smarter than us? No. Do they have more production capacity? No. Are they ahead of us? Yes. Why? Because it's like full-throated endorsement of building out"