The Good Thing Podcast

Why Infrastructure Should Just Work with Sam Lambert

August 1, 2025
with Sam Lambert
Hosted by Jens Neuse & Stefan Avram
Directed by Jacob Javor

PlanetScale CEO Sam Lambert joins The Good Thing to talk scale, databases, and why boring infrastructure is not a problem but a goal.

About Sam Lambert

CEO, PlanetScale

Sam Lambert is the CEO of PlanetScale, the database platform built on Vitess and now expanding into Postgres. Previously, he was VP of Engineering at GitHub and worked on internet-scale infrastructure at Meta.

Visit Sam Lambert's website →

TL;DR

PlanetScale CEO Sam Lambert shares why databases need to be invisible, how small teams can outperform massive ones, and what it really takes to run infrastructure at scale. From Postgres to Kubernetes to capital efficiency, Sam keeps coming back to one idea: the best tech just works.

Postgres is the next MySQL

PlanetScale built its reputation on MySQL and Vitess. But now it’s shipping Postgres too. Sam says the ecosystem is maturing fast.

Postgres is probably going to go through the same era that MySQL went through from like 2010 to 2020.

That shift opens up new surface area for PlanetScale, which now positions itself as “the database company,” not just a MySQL platform.

Reliability by design, not by headcount

PlanetScale runs with just 55 people, including two in marketing and three in sales. Its database fleet is self-healing by default. “There’s no sysadmin-style people,” Sam said. “It’s all writing automation that manages the fleet continually.”

If you build good infrastructure, you shouldn’t need to throw tons of headcount at it.

The team’s confidence comes from deep experience. “Pretty much every engineer at PlanetScale could be a principal engineer at any other company,” Sam added.

Profitability beats vanity metrics

Sam doesn’t put much stock in user count graphs or burn-to-grow strategies. He prefers durable growth—and he’s clear-eyed about what that means.

It’s not a business until it makes money. When you see more money going into the bank than leaves—that is a business.

PlanetScale has been profitable, with surplus quarters and a lean hiring model. Sam’s take: “You’re not beating Domino’s because you’re giving away free pizza outside Domino’s.”

“It’s all theory until it works at extreme scale.”

For Sam, infrastructure doesn’t count unless it holds up under pressure. That’s why he still loves databases—and why he doesn’t need flashy dashboards to appreciate what PlanetScale does.

I watched a customer spin up a cluster… 500 servers orchestrate, spin up, report for duty, and immediately act in unison to serve a database through one connection string.

AI is adding fuel, not replacing fundamentals Sam doesn’t think AI changes everything, but it is changing demand. “AI agents are creating new usage patterns and new things at scale,” he said. “It’s just brought more load, more demand, more concurrency.”

PlanetScale even counts Cursor as a customer and uses Cursor internally. “It’s really fun to be in the loop,” Sam said.


This episode was directed by Jacob Javor. Transcript lightly edited for clarity and flow.

Frequently Asked Questions

How does PlanetScale achieve high availability on metal?

By running ephemeral machines with NVMe drives inside AWS and GCP, PlanetScale avoids the latency of disaggregated storage while maintaining HA through semi-synchronous replication.

What makes PlanetScale competitive despite its size?

Sam Lambert points to expertise and operational maturity. PlanetScale runs with just 55 people. This including 4 on infrastructure and 10 on Vitess—and operates with no sysadmins, relying instead on automation.

Never miss an episode

Subscribe to The Good Thing to get notified when new episodes drop.