Rent or Own? How the "Rug Pull" Era is Pushing Developers Toward Source-Available Software

From “rug pulls” to corporate warfare, the rules of software ownership are being rewritten. But a new model—Post-SaaS—might finally offer a way out.

For decades, the deal was simple: developers gave their time, and companies gave their code. It was an unwritten social contract built on trust. Then, in a few short months, the contract was shredded.

First came the tremors. In August of 2023, HashiCorp announced that Terraform—the industry standard for infrastructure-as-code—was switching to a “Business Source License,” effectively walling off competitors. The aftershocks followed in March 2024, when Redis, the database powering a considerable chunk of the modern web, abandoned its open-source roots in favor of more restrictive terms.

To the C-suites, these were necessary pivots to protect revenue from cloud giants like AWS. But to the millions of engineers who had built their careers and stacks on these tools, it felt like something else entirely: a rug pull.

The premise of these shifts was that the “Open Core” model was broken—that you couldn’t build a profitable business by giving away the recipes. But this reactionary move misses a fundamental truth about the modern software economy. By trying to lock down their code, these companies didn’t just lose the moral high ground; they inadvertently proved that in 2026, the code itself is no longer the asset. The community is.

The Revolt: Why You Can’t Close the Barn Door

When a company closes a previously open project, they don’t just lose users; they create a martyr.

The immediate reaction to the Terraform and Redis announcements wasn’t just anger—it was action. The Linux Foundation stepped in, backing “OpenTofu” (a fork of Terraform) and “Valkey” (a fork of Redis). Almost overnight, the original companies found themselves competing against free, community-driven versions of their own products.

As Madelyn Olson, a core Redis maintainer who left to build Valkey, put it: “I worked on open source Redis for six years... By forming Valkey, contributors can pick up where we left off and continue true open source development.”

The lesson here is stark: You cannot retroactively close a community-built project without destroying your reputation. The talent leaves, the momentum shifts, and the “rug pull” strategy often backfires, creating a new competitor with the moral high ground.

The Platform Paradox: Why the “Checkroom” Makes More Than the Coat

If the “Rug Pull” is a desperate attempt to monetize the asset, the “Platform” model proves you don’t need to own the asset to monetize it. You need to be the best place to keep it.

The clearest example of this is Hugging Face. Often described as the “GitHub of AI,” Hugging Face hosts over one million models, datasets, and demos—almost all of them open source and free to download. By the logic of the “Rug Pull” CEOs, this should be a disaster. Why would anyone pay Hugging Face when they can download the Llama 3 weights and run them locally?

The answer lies in the friction of modern infrastructure. Hugging Face generated over $70 million in revenue (2023), not by gating access to the algorithms, but by selling the “compute” and “enterprise security” required to run them.

They understood a fundamental truth about developers: we are lazy and busy. We could spin up our own AWS instances, configure the CUDA drivers, and secure the endpoints—or we could pay Hugging Face $0.50 an hour to click a single button labeled “Deploy”.

This is the Platform Moat. While HashiCorp and Redis were busy building legal fences around their code, Hugging Face was building a toll road. They realized that in an era of abundant open-source software, the scarce resource isn’t the code; it’s the convenience.

"By trying to lock down their code, these companies didn't just lose the moral high ground; they inadvertently proved that in 2026, the code itself is no longer the asset. The community is."

The Weapon: Meta’s Scorched Earth Strategy

While Hugging Face proves you can build a business on top of open source, Meta proves you can use open source to burn a competitor’s business to the ground. This is the strategy known in economics as “Commoditizing the Complement.”

For OpenAI and Google, the AI model is the product. They spend billions training GPTs and Gemini, intending to rent access to them. Their entire business model relies on the model being a scarce, proprietary secret.

Enter Meta. By releasing Llama—a state-of-the-art LLM—for free, Mark Zuckerberg isn’t just being altruistic; he is devaluing the core product of his rivals. If developers can get 95% of GPT -4’s performance for $0 by using Llama, the market price for “intelligence” drops toward zero.

This is a defensive play ripped straight from the 2000s playbook. Just as Google released Android for free to prevent Microsoft and Apple from owning the mobile internet, Meta is releasing Llama to prevent OpenAI from owning the AI internet.

For the developer, this is a windfall. We get enterprise-grade tools without the enterprise price tag. But let’s be clear about the dynamic: we aren’t being given a gift; we are being handed ammunition in a war between giants. Meta’s bet is simple: if everyone builds on Llama, the ecosystem locks into their standards (PyTorch, etc.), and the “walled gardens” of closed AI find themselves guarding an empty castle.

Author’s disclaimer: Meta’s AI business strategy and business strategy of their other products like Facebook, Instagram, and WhatsApp are vastly different, which I, along with many others, try to avoid—but that is a story for another time.

Conclusion: The Post-SaaS Reformation

If the “Rug Pulls” of 2024 taught us that we can’t trust corporations to keep their code open, and the “AI Wars” taught us that open source is often just a weapon for giants, where does that leave the rest of us?

It leaves us looking for a third way—one that rejects both the “Rental Economy” of SaaS and the “Rug Pull Risk” of open core.

Enter the “ONCE” philosophy, championed by 37signals (the creators of Rails). With the launch of Campfire and Writebook, they introduced a model that feels radical simply because it is retro: You pay once. You install it. You own it.

David Heinemeier Hansson calls this the “Post-SaaS” era. The license isn’t Open Source (you can’t resell it), but it is Source Available. More importantly, it is irrevocable. Once you download the code to your server, no board of directors can change the terms. No acquisition by a competitor can shut it down.

As Hansson puts it: “SaaS is the ultimate trap. You rent your tools, you rent your data, and the landlord can raise the rent whenever they want. ONCE is about returning to software you actually own.”

This is the lesson for the next decade of the software business. The “Moat” is no longer the code—it’s the trust. Developers are tired of building on quicksand. Whether it’s through the “Platform Model” of Hugging Face or the “Ownership Model” of ONCE, the winning companies of 2026 will be the ones that sign a new contract with their users: We don’t want to lock you in. We want to be so good you don’t want to leave.

Are we entering a period in which the most profitable software is free and open for personal use? The time will tell.