The Fork That Actually Stuck
In August 2023, HashiCorp made a decision that fractured the infrastructure-as-code community. They switched Terraform’s license from Mozilla Public License 2.0 to the Business Source License, a move that essentially told open-source purists their free ride was ending. The Linux Foundation didn’t waste time. Within months, OpenTofu emerged as a community-backed fork, and here we are in early 2026 watching what happens when someone calls a corporation’s bluff and actually builds an alternative.
This isn’t theoretical anymore. OpenTofu has crossed into genuine competitive territory. The project now sits at over 23,000 GitHub stars with more than 4 million weekly downloads as of January 2026. That’s a 166% jump from the 1.5 million downloads recorded at launch. These aren’t vanity metrics. When infrastructure teams download a tool 4 million times per week, they’re voting with their CI/CD pipelines.
What makes this moment interesting isn’t the fork itself. Open-source forks happen constantly. What matters is that OpenTofu has achieved something forks rarely do: it’s become the preferred choice for a meaningful portion of the market.
State Encryption: The Feature That Matters
Let’s talk about what separated OpenTofu 1.9 from Terraform 1.10 on the technical level, because this is where the fork stops being symbolic and starts being real. OpenTofu 1.9 introduced native end-to-end state encryption. Terraform’s open-source tier doesn’t offer it.
State files in Terraform are where the magic happens and where your secrets live. If you’re managing infrastructure at any serious scale, your state file contains database passwords, API keys, encryption keys, and enough sensitive information to make a security auditor weep. Terraform Cloud offers encryption at rest, but that’s the paid tier. The open-source version? You’re responsible for securing your own state backend. Most teams use S3 with server-side encryption, which works fine until it doesn’t. OpenTofu took the harder path and built encryption into the tool itself.
This matters more than it might seem. If you’re a mid-market company that can’t justify Terraform Cloud costs, or if you’re philosophically opposed to relying on HashiCorp’s managed service, OpenTofu’s native encryption closes a gap that’s been festering since Terraform became popular. You get enterprise-grade state management without the enterprise pricing model.
The IBM Era and What It Means for Terraform’s Direction
Here’s the thing nobody wants to say out loud but everyone’s thinking: HashiCorp’s acquisition by IBM in April 2024 for 6.4 billion dollars changed the energy. IBM doesn’t acquire companies to maintain them as experimental playgrounds. IBM acquires companies to fold them into enterprise platforms and extract revenue.
Terraform 1.10 has been incrementally good. Stability improvements. Performance tweaks. The releases work. But they feel managed rather than driven. The community’s watching Terraform 1.10 and asking a question that’s hard to take back once you ask it: are we going to see the same velocity of innovation we saw pre-acquisition? When your corporate parent has quarterly earnings calls to worry about, open-source innovation starts looking like a cost center rather than a strategic advantage.
OpenTofu doesn’t have that problem. It’s got Linux Foundation governance, which means decision-making happens in the open and nobody’s quarterly revenue depends on pushing certain features over others. The CNCF Technical Oversight Committee accepted OpenTofu as a sandbox project in early 2025, the same governance pathway that Kubernetes and Prometheus used to build enterprise credibility. That’s not luck. That’s institutional legitimacy.
Migration Is Happening Faster Than Anyone Expected
The Pulumi State of Infrastructure as Code survey from January 2026 surveyed 1,200 platform engineers and found that 31% had already migrated at least one environment from Terraform to OpenTofu. That’s a three-fold increase from 11% the year prior. In enterprise decision-making, 20-point swings in adoption within 12 months don’t happen by accident. They happen when enough people conclude that the alternative is worth the switching costs.
Switching infrastructure-as-code tools is not trivial. You’ve got state migration, provider compatibility validation, team retraining, and the lurking fear that you’ll discover an edge case in production at 2 AM. Teams don’t make these moves lightly. The fact that 31% have already migrated at least one environment tells you something: the perceived benefits are outweighing the switching friction.
What makes this migration wave meaningful is the directionality. These aren’t teams forcing OpenTofu on themselves as some philosophical purity test. These are teams running a cost-benefit calculation and concluding that OpenTofu solves their actual problems better than Terraform does right now.
What This Means for Your Infrastructure Decisions
If you’re building new infrastructure or refreshing your tooling strategy, OpenTofu deserves serious consideration. The documentation is solid. The provider ecosystem is growing. The community is active and genuinely invested in making this work. OpenTofu official documentation and changelog is worth reviewing, and the Linux Foundation OpenTofu project page lays out the governance structure backing the project.
If you’re deep in the Terraform ecosystem already, there’s no need to panic. Terraform still works. It’s still supported. But you should probably start thinking about what an OpenTofu migration path looks like for your organization, because the fork just became real competition, and competition is usually good for the ecosystem. When vendors have to earn your loyalty instead of assuming it, you win.
The infrastructure-as-code landscape has shifted. The fork isn’t theoretical anymore. It’s 23,000 stars and 4 million weekly downloads. It’s 31% of platform engineers trying it in production. It’s CNCF legitimacy and native encryption and Linux Foundation backing. If you’ve been watching from the sidelines, now’s the time to form an actual opinion.
What’s your take? Have you tested OpenTofu in a sandbox environment yet? I’d be genuinely curious what friction points you’ve hit and whether the migration story actually holds together when you stress-test it against your real infrastructure.