SaaG (Software as a Graveyard)

Most SaaS companies raised around 2021 valuations are already dead. They just don’t know it yet.

If you know me, you already know I’m a Naval fanboy. Hearing what I was thinking out loud in his last podcast felt scary for me:

“Pure software is uninvestable. Full stop.”

I don’t think he was hedging. He was basically describing a regime change that most founders around me are still pretending is a phase.

I mean, the pitch was always the same… Building this is hard. We have a team. We have a head start. The moat is execution.

Today, execution stopped being scarce.

Two devs with the right tooling rebuild the working surface of most B2B SaaS in a quarter. Or a non-developer “idea guy” like me is able to build a mini B2C SaaS overnight. Not a prototype. Real architecture with room to scale. What’s left is integrations, enterprise sales, compliance posture. That’s friction. Friction is not a moat. Friction gets compressed every six months as the tooling and agents get better.

The cycle is already pretty much visible. Notion ate a generation of wikis and doc tools (and much more, I love you Notion) that raised real money on the premise that a flexible block editor was hard. Linear did the same to Jira’s mid market by being faster and better designed, not by inventing a category. Linear is now the incumbent watching teams ship a competitive issue tracker in a few weekends.

Each generation of winners has less time on top before the next one arrives with a tenth of the headcount.

Mailchimp sold to Intuit for 12 billion in 2021 on twenty years of SMB email infra. Yes. The brand and merchant relationships are still real. But I don’t think software underneath is an asset anymore.

The honest test for a founder is one question. What does an AI native team of three, fully funded, ship against you in six months? If the answer is most of it… Do you believe you can survive? What survives this is narrower than most of us want to admit.

One thing is clear though. Distribution survives. This is also the biggest challenge we have at my company today. We are able to push out great products super quickly, thanks to Claude and our brilliant people, yet distribution is something that always came as an afterthought. We were wrong. The companies that own the relationship, the inbox, the channel, the trust, are durable in a way the product cannot be. Product is downstream of attention. Most founders treat marketing as something that happens after the build is done. They have it backwards. We had it backwards. Yes, network effects survive, but only the real ones. If the product is meaningfully better at a hundred thousand users than at a hundred, you have something. If it works the same, you have a feature. Feature sets depreciate.

Proprietary data survives when it actually compounds. Bloomberg is a data company with a terminal stapled on. Tesla’s autonomy stack is a data business with wheels. If users “generate” something with every interaction that a competitor cannot buy, scrape, or synthesize. Trust and regulatory depth also survive. Stripe’s moat is not the API (anymore). It is a decade of compliance, banking relationships, fraud handling, and dispute infrastructure that nobody wants to rebuild. It’s the distribution, ease of use, and “habits” of existing customers that took ages and lots of headaches to build. Vertical specialists who own the workflow, the regulator, and the audit trail of one industry are in a better position than horizontals defending everything at once.

I think hardware will still survive the longest. Atoms commoditize slowest. Nothing about manufacturing or physics is getting cheaper at the rate code is (well, until we reach AGI and that lovely AGI does unexpected scientific breakthroughs that will drastically reduce the cost of manufacturing and but we’re also able to 3D print nuclear weapons?)

Long story short, the ceiling on what one person can build is just removed.

One person companies operating with the velocity of fifty. The founder’s vision compiled directly to product, without the dilution that happens when you explain it to twelve other people who each want to leave fingerprints on it.

Echoing Naval again, he says there are three options from where most founders are sitting — I almost fully agree with them so listing them below with my additional Turkish spice:

1. Decide this is hype or a “bubble”. Keep building like nothing changed. Find out at the next round that comps moved without you. This is the popular choice and it will produce the bulk of the layoff posts in 2027.
2. Realize it late. Cut under stress and make stupid decisions. Pivot without leverage. The founders who get destroyed by this transition are not the ones who saw it coming. They are the ones who waited until burn forced the decision.
3. Or… Take the next 1–2 years you have seriously. Audit the moat honestly, not at the offsite. Build distribution before you need it. Go vertical deep where horizontals are overextended. Own the data or the workflow or the regulator or the physical layer. Anything that does not sit on top of a model someone else trains.

The window is open right now. It will probably not stay open.

The companies that come out of this are the ones whose founders read the room early and repositioned while they still had the runway and the leverage to do it cleanly.

The rest will go on podcasts explaining how it happened so fast.