←back to thread

Rocketships and Slingshots

(postround.substack.com)
42 points juecd | 1 comments | 04 Sep 25 21:06 UTC | HN request time: 0.202s | source
Show context
alphazard ◴[04 Sep 25 22:37 UTC] No.45133025[source]▶
No examples included of "hard earned technical moats". I'm a bit skeptical that these really exist, other than in a few extreme cases. A true technical moat requires secrecy and control of the talent pool. SpaceX comes to mind, Nvidia might be another example.

I don't even think the current AI companies really have technical moats. It seems like the differentiator is just how much money they have to throw at compute.

Most moats are actually regulatory, ranging from copyright and patents, to anti-competitive regulation, to explicit wealth transfers from taxpayers to the company.

replies(3): >>45133155 #>>45133772 #>>45140525 #
akharris ◴[04 Sep 25 22:52 UTC] No.45133155[source]▶
My list wasn't meant to exhaustive. Honestly not much of a list, just a few examples. SpaceX and Nvidia are both examples of deep technical moats, as are some of the others mentioned by commenters.

I agree with you that many moats are regulatory, but disagree that most moats are. Networks are powerful moats, customer lockin can work well, brand is a strong moat (at least for a time), speed and culture can function as moats.

Fwiw, I don't think any moat is ultimately perfect. Moats themselves are not meant to permanently stop an attacker - they are delaying tactics at best. Companies that stay ahead continuously re-invent and rebuild their differentiation and defenses.

replies(2): >>45133388 #>>45135174 #
1. alphazard ◴[04 Sep 25 23:26 UTC] No.45133388[source]▶
Relationships with suppliers and customers are what gives any business most of its value.

I don't consider that to be part of a moat because you don't have to cross it before you begin competing with the business. Establishing suppliers and customers is just competing with whatever businesses would also want those deals.

A technical moat is knowledge that you have to have (maybe across multiple employees) before you can even begin to start competing. It's rare to come across any technology that couldn't be cloned by a few smart engineers in 6 months. Much more likely that you aren't allowed to build or sell the technology, or you don't have enough capital, or no one will sell you the parts, or it's illegal to start the business in the first place.