Ask HN: What are tech's largest moats?

[miguelrochefort]

Let's imagine you're building a company that intends to compete with FAANG.

Your company will be expected to have most of these:

* AI Research Lab (FAIR, Brain, MLR, Research)

* ML Framework (PyTorch, TensorFlow, CNTK, Core ML)

* Web Framework (React, Angular, Blazor)

* Browser (Chrome, Safari, Edge)

* Smart Speaker (Nest, HomePod, Echo)

* Smartphone (iPhone, Pixel, Galaxy, Fire Phone, Lumia)

* VC Fund (GV, M12, Apple, Alexa Fund)

* Virtual Assistant (Siri, Alexa, Assistant, Bixsby, Cortana)

* Personal Cloud Storage (iCloud, Drive, OneDrive)

* Typeface (Roboto, San Francisco, Segoe)

* Personal Video Chat (FaceTime, Duo, Skype, Messenger)

* Work Video Chat (Meet, Teams, Chime)

* Personal Chat (iMessage, Hangouts, Skype, Messenger, WhatsApp)

* Chip (Silicon, Tensor, Pluton, Graviton)

* App Store (App Store, Play Store, Microsoft Store)

* IDE (Visual Studio, Android Studio, Xcode, Cloud9, Nuclide)

* Search (Google, Bing)

* True Wireless Earbuds (AirPods, Pixel Buds, Galaxy Buds, Echo Buds, Surface Buds)

* Email (Gmail, Outlook)

* Programming Language (C#, Swift, Go, Hack)

* In-Car Entertainment (CarPlay, Android Auto)

* Design (Human Interface Guidelines, Material, Fluent)

* Maps (Google Maps, Bing Maps, Apple Maps)

* Notebook Hardware (MacBook, Surface, Pixel Book, Galaxy Book)

* Desktop OS (Windows, macOS, Chrome OS, DeX)

* Wearable (Apple Watch, Pixel Watch, Samsung Watch, Halo, Band)

* Health (Apple Health, Google Fit, Samsung Health, HealthVault, Halo)

* Smart TV (Apple TV, Chromecast, Fire TV)

* Office Productivity (Office, Docs, iWork)

* and much more...

Some of these are easy to reproduce, through licensing, partnerships, acquisitions, reverse-engineering, or open source forking. For example, it is possible for small companies to design their own typeface, white-label their brand of true wireless earbuds, or develop a basic chat application.

On the other hand, some are incredibly difficult to reproduce, due to network effect, lack of data, or R&D costs. For example, designing your own chip, developing your own office productivity suite, obtaining movie streaming rights from all major film producers, or building a digital map of all the streets in the world, might only be achievable by a handful of tech giants. As such, these would be considered to be some of tech's largest moats.

What are some of the largest moats in tech? Are there technologies out there that a company with 5 billion dollar and 5 years would struggle to replicate?

[mikewarot]

Software has essentially zero marginal cost. There's nothing stopping you, or any of us here, from writing another social network site. Management is the really hard thing to get right, if you want to grow to market dominance.

At its core, only Google's search engine is something of a moat. You have to crawl a sufficiently large chunk of the Web to yield good results.

You correctly point out Semiconductor Manufacturing as having a large cost of entry, which is true of any business that deals in atoms instead of bits.

Heck, just to set up a shop in your garage to make gears would probably set you back $10,000. But you could write the code to control the CNC machine, etc.. on a $100 netbook.

[tomcam → mikewarot]

> There's nothing stopping you, or any of us here, from writing another social network site. Management is the really hard thing to get right, if you want to grow to market dominance.

What? No. It’s the network effect: getting enough users early on to achieve critical mass

[mikewarot → tomcam]

If the network is managed well, they'll keep users, and grow through word of mouth...if not...it'll tank.

[AveburySutton → mikewarot]

>There's nothing stopping you, or any of us here, from writing another social network site.

Software patents has entered the chat...

[jstx1]

> Let's imagine you're building a company that intends to compete with FAANG. Your company will be expected to have most of these

Not really - you'll need most of them if you want to imitate the big tech companies but you can compete in only 1 category, or in a small number of categories. You don't need all of them.

Also, many of these are cost centers - research, open source languages, frameworks and IDEs, browsers etc. Competition doesn't work the same way in that space.

The one that's obviously missing from your list and probably has the highest barrier for entry is cloud - very tough market, a few established players, huge upfront cost, lots of expertise required to do well and lots of potential downside if you don't do it well.

[kingkongjaffa]

Microsoft excel driving an astonishingly large number of business processes.

[webmaven → kingkongjaffa]

> Microsoft excel driving an astonishingly large number of business processes.

I'm not sure I agree. Google obviously felt that they needed an office suite (including a spreadsheet) in order to avoid being frozen out of some markets (so they bought one), but none of the other Big Tech players felt the same way, nor does the lack of one seem to have actually limited their growth, and the current crop of low-code and no-code tool entrants don't seem to be predicated on Excel as a starting point.

[I_DRINK_KOOLAID]

Windows GUI you just can't attack it.

Billions of kids grew up with the start button as a gateway to their favorite game, and billions of adults sees it as the gateway to their productivity tool

[usrn → I_DRINK_KOOLAID]

Maybe it's because the last version of Windows I had installed was XP SP1 but I can't find things in the modern Windows start menu.

[I_DRINK_KOOLAID → usrn]

It's the reason why IIRC Server still has the old Start button which emerged in Windows 95 . Consumers version are plagued by the new menu but you can cure it with a free software called "classic start menu" or something like that.

[markus_zhang]

Political-Financial connections.