As a community, and as individuals/businesses, we have to make some big decisions about what to open source. The Tailwind layoffs and PR refusal force us to face reality.

Why Different Models Work for Different Businesses

Open source isn't one strategy. It's a tool that works differently depending on what you're building and why.

The Infrastructure Play: Supabase

What they open sourced: Client libraries, dev tools, local dev environment
What they kept: Managed infrastructure, scaling, operations

Why it works:

Easy to test locally and build prototypes
Real production use needs reliability and scale
Self-hosting Postgres is technically possible but operationally expensive
Most teams would rather pay than manage infrastructure

The business model: Open source drives adoption, infrastructure drives revenue. You try it for free, you scale with them.

Pros:

Maximum developer adoption (low friction to start)
Community contributions improve the product
Self-hosting option builds trust
Natural upgrade path from free to paid

Cons:

Need significant capital to run infrastructure
Cloud providers could compete on price
Hard to monetize small/mid-market

The Framework + Tooling Play: LangChain

What they open sourced: The entire framework
What they sell: LangSmith (observability, debugging, monitoring)

Why it works:

Framework needs to be everywhere to be standard
Real AI applications need debugging and monitoring
Community makes the framework better than any fork could
Hard to copy their hosted tooling

The business model: Give away the building blocks, charge for making them work in production.

Pros:

Framework adoption drives tooling demand
Community contributions strengthen moat
Clear separation between free and paid value
Network effects (more users = better framework)

Cons:

Framework could be forked if tooling gets too expensive
Competition from observability companies
Need to keep innovating on tooling side

What they open source: SDKs, examples, templates, deployment tools

Why it works:

Makes it easy to use their compute
More usage = more revenue
Open tools reduce friction to adoption
Can't really "steal" the infrastructure

The business model: Open source is marketing. Revenue is compute time.

Pros:

Every user of the open source tool is a potential customer
Community creates examples and tutorials (free marketing)
Open source builds trust in a commodity market
Clear value exchange (tools free, compute paid)

Cons:

Competitors can use your tools with their infrastructure
Race to bottom on compute pricing
Hard to stand out beyond price

The Problem: Tailwind CSS

What they open sourced: The entire CSS framework
What happened: Companies packaged it and sold it

Why it got complicated:

All the value is in the code itself
No natural "hosted service" layer
Easy to package and resell
Hard to add premium features without making the base version feel incomplete

The mismatch: They open sourced a finished product, not a building block for a service.

What they're learning: Sometimes being "too good" at open source means there's nothing left to monetize.

The Pattern

Looking across these models, you see a clear pattern:

Open source works when:

The open source part drives usage of a paid service
Self-hosting is technically possible but operationally expensive
Community contributions make the product better
There's a natural split between "building blocks" and "production infrastructure"

Open source struggles when:

All the value is in the code itself
Easy to package and resell
No natural service layer to monetize
Community contributions don't add significant value

Now Add AI to the Equation

AI changes these things in two ways:

1. Training Data Concerns

Your open source code/docs can become training data for competing products. This is new territory.

For frameworks: Probably fine. If someone trains an AI on LangChain docs, it just drives more LangChain adoption.

For finished products: Riskier. If someone trains an AI on your complete implementation, they might copy your approach.

For domain expertise: Here's the interesting one. If your value is in accumulated knowledge (compliance frameworks, best practices, specialized techniques), AI can potentially extract and copy that.

2. AI-Native Products Need Open Documentation

Products that work with AI agents (dev tools, workflows, integrations) need to be AI-readable to be useful. But making everything AI-readable makes it easier to train on.

The tension: You want AI to understand your product well enough to use it. You don't want AI to understand it well enough to replace it.

Four Strategic Approaches (Pros/Cons)

Approach 1: Full Open Source + Hosted Services

Model: Everything open, revenue from managed services
Examples: Supabase, Sentry, GitLab, Plausible

Pros:

Maximum adoption and community trust
Self-hosting option builds credibility
Community contributions improve product
Clear value proposition (convenience + scale)

Cons:

Need capital to build infrastructure
Cloud providers can compete
Need ongoing ways to stand out on service layer
Can't really "take back" what you open sourced

Best for: Products where the hard part is infrastructure, not code

Approach 2: Open Framework + Paid Tooling

Model: Core product free, productivity tools paid
Examples: LangChain, Terraform, Docker

Pros:

Framework adoption drives tooling demand
Community strengthens the framework
Natural upgrade path for growing teams
Hard to copy the ecosystem

Cons:

Need to keep innovating on tooling
Risk of framework fork if pricing is wrong
Tooling competitors can build on your framework
Need both framework AND tooling expertise

Best for: Platforms where adoption matters more than direct revenue from core product

Approach 3: Open Source as Marketing

Model: Open source the interface, charge for the service
Examples: Compute providers (Modal, RunPod), AI API providers (Anthropic SDKs)

Pros:

Every user is a potential customer
Reduces friction to adoption
Community creates examples (free marketing)
Can't really "steal" the underlying service

Cons:

Competitors can use your tools
Commodity service (compete on price/quality)
Need continuous service advantage
Open source doesn't create moat

Best for: Service businesses where open source tools drive usage

Approach 4: Source Available (Not True Open Source)

Model: Code is visible but has usage restrictions
Examples: HashiCorp (post-license change), Redis (SSPL), Elastic

Pros:

Transparency without full vulnerability
Can prevent direct competition/reselling
Maintain control over commercial use
Source visibility builds some trust

Cons:

Community backlash ("not real open source")
Hard to build true community
Confusing licensing terms
May fracture ecosystem (old vs new license)

Best for: Products worried about cloud provider competition, but this is playing with fire

What This Means for You

If you're thinking about open source strategy:

Ask yourself:

What's the hard part of what you're building? (Code? Infrastructure? Expertise? Network effects?)
Can the hard part be commoditized by AI training on your open source materials?
Is there a natural service layer to monetize? (Hosting? Tooling? Support? Updates?)
Does community contribution make your product better or just give away what makes you different?
Can you build/improve faster than someone could fork?

Red flags that open source might not work:

All your value is in the code itself
Easy to package and resell
No natural ongoing service
What makes you different is technique, not execution
Small market where every customer counts

Green flags that open source could work:

Infrastructure/operations is the hard part
Community contributions add real value
Natural freemium or usage-based model
Open source drives adoption of paid services
Large market where awareness matters

The Questions No One Has Answered Yet

Here's what I'm watching:

Will HashiCorp's license change work? They went from open to restricted. Community is mad. Will they lose developers or gain revenue?

What happens when AI tooling becomes a commodity? LangChain's model works because LangSmith is hard to copy. What happens when AI makes it easy?

Can Tailwind find a new model? They built something too good to monetize traditionally. What's their next move?

Do "open source but not really" licenses work long-term? Source-available licenses feel like trying to have it both ways. Does the market accept them?

The Real Answer

There's no universal answer. It depends on:

What you're building
Who you're building for
Where the value lives
How fast the market moves
Whether AI can copy what makes you different

The companies that win are the ones who understand where their actual moat is and structure their open source strategy around protecting it while still capturing the benefits of community and adoption.

The companies that lose are the ones who:

Open source everything and realize too late there's no moat
Keep everything closed and miss network effects
Pick the wrong thing to monetize
Change strategy too late or too abruptly

What's your take? What model are you using? What's working?

Open Source Business Models in the Age of AI: What Works, What Doesn't, and Why

The Tailwind layoffs force us to rethink what we open source and why