Best European Alternatives to Terraform
Looking for a European alternative to Terraform? HashiCorp relicensed Terraform under the Business Source Licence in 2023 and was acquired by IBM in 2024 — neither breaks existing usage, but together they made a lot of platform teams uncomfortable betting a decade of infrastructure code on one US vendor.
OpenTofu is the community answer, and Europe supplies most of the orchestration layer around it — Polish, German and British companies running the automation that applies your plans.
How we rank these tools — 4-step process
-
1
European ownership, verified
The company is headquartered and incorporated in the EU, EEA or Switzerland, and processes customer data in Europe. A US parent company disqualifies a tool from this page regardless of where its servers are.
-
2
Category fit and hands-on review
What the tool actually does, who it suits, and where it falls short — checked against the vendor’s own documentation, changelog and pricing page rather than its marketing copy.
-
3
Compliance and pricing check
GDPR posture, hosting location and the prices quoted on this page are verified against the vendor’s public pricing before publication, and re-checked when we revisit the category.
-
4
Position on this page
Placement on this page can be paid, and that can affect which tools appear here and the order they appear in. It never buys a good review: a tool that fails the checks above is not here at any price, and payment does not change the shortcomings we write about. A vendor can ask us to correct a factual error — not to remove a criticism.
Vendors can pay for visibility on this page. It never changes what an entry says about a product, including the criticism, and we earn nothing when you click through to a vendor. Paid placement can affect which tools appear here and the order they appear in. Editorial policy
7 European Alternatives to Terraform
OpenTofu
The community fork of Terraform, under the Linux Foundation
Spacelift
Managed CI/CD for OpenTofu, Terraform, Pulumi and Ansible
NixOS
Declarative, reproducible system configuration
Buddy Works
Pipeline automation with an unusually approachable UI
Terramate
Orchestration and code generation for large IaC estates
Canonical Juju
Model-driven operators for cloud and Kubernetes
Digger
Runs IaC plans inside your own CI instead of a vendor's
Key takeaways
- Terraform stopped being open source on 10 August 2023; the LICENSE file now names International Business Machines Corporation as Licensor under BSL 1.1.
- IBM completed the HashiCorp acquisition on 27 February 2025 for $6.4 billion, so the licence and the roadmap now sit with IBM.
- The additional use grant permits internal production use, so most teams are not in breach — the risk is that the terms can change again, not that you are breaking them now.
- This is the one migration on this site that is genuinely easy: OpenTofu accepts the same HCL and the same state file, so switching is changing which binary runs.
- OpenTofu is not European. It is a Linux Foundation project with no single jurisdiction, and on this page governance is the point rather than geography.
Why people leave Terraform
Terraform is not a service that holds your data. It is a binary that reads your files and calls cloud APIs from your own machine or your own CI runner, so the argument that carries most comparisons on this site — an American company is holding something of yours — does not apply unless you are paying for HCP Terraform. Pretending otherwise would be the fastest way to get this page wrong.
What changed is the licence. On 10 August 2023 HashiCorp moved Terraform from MPL 2.0 to the Business Source License 1.1, and the LICENSE file in the repository today names International Business Machines Corporation as the Licensor.
The Change Date is four years from publication of each version, the Change License is MPL 2.0, and the Additional Use Grant permits production use provided you do not offer the work hosted or embedded in competition with IBM's paid versions. IBM completed the $6.4 billion acquisition of HashiCorp on 27 February 2025.
For the overwhelming majority of teams that grant covers what they do, so nobody is breaking a rule by running terraform apply against their own accounts. The reason to move is narrower and harder to dismiss: infrastructure code is the most expensive code an organisation owns to rewrite, and its terms of use are now set unilaterally by a single licensor who has already changed them once.
- The licensor is now IBM Read the LICENSE file rather than the marketing: Business Source License 1.1, Licensor International Business Machines Corporation, Change Date four years from publication, Change License MPL 2.0. Nothing about that is hidden and nothing about it is illegal. It simply means the terms under which your production infrastructure is described are set by one company, and the company setting them is not the one that set them when you wrote the code.
- The additional use grant is a boundary you cannot see The grant allows internal production use and stops at offering the work hosted or embedded in a way that competes with IBM's paid products. If you are a platform team running your own estate, you are comfortably inside it. If you sell anything that provisions infrastructure for customers, you are near a line whose position is described in prose rather than in code, and which HashiCorp already needed a clarification round in October 2023 to explain.
- European residency exists but is not on the price list HCP Terraform publishes per-resource pricing — Essentials from $0.10 per resource per month, Standard $0.47, Premium $0.99 — and lists a separate offering, IBM HCP Terraform Europe, for enterprises that need geographic service locality. Its price is "custom". So EU locality is available and it cannot be bought self-serve: you get it by opening a procurement conversation, not by choosing a region in a dropdown.
- The state file is the real asset Terraform state records the identity of every resource you own, and in plain form it contains secrets. Whoever holds that file holds an inventory of your estate. That is the question worth asking about any replacement, and it is why two of the tools here refuse to hold it at all and run inside the CI system you already secure.
What you have to replace, not just match
Terraform is two products with one name, and the switch is completely different depending on which one you are leaving.
The first is the CLI and the language: HCL, providers, modules, the state file, terraform plan and terraform apply. The second is HCP Terraform, formerly Terraform Cloud, which adds remote runs, state storage, policy enforcement, approvals and a private module registry, billed per resource under management.
If you are leaving the CLI, OpenTofu is the whole answer and the migration is a binary swap on the same state file. If you are leaving HCP Terraform, the engine is not the question at all — you need somewhere for runs to happen and something to hold state, and Spacelift, Digger, Terramate and Buddy Works are four different opinions about where that should be. Most teams turn out to be leaving the second and not the first.
The alternatives compared
| Position | Tool | Headquarters | Pricing | Jurisdiction |
|---|---|---|---|---|
| #1 | OpenTofu | Distributed | Free and open source | Vendor-neutral; no single jurisdiction |
| #2 | Spacelift | Warsaw, Poland | Free tier / Cloud from about $400/month / Self-Hosted on request | EU (Poland) engineering, US-registered entity |
| #3 | NixOS | Utrecht, Netherlands | Free and open source | EU (Netherlands) |
| #4 | Buddy Works | Poland | Paid plans plus Bring Your Own Cloud | EU (Poland) |
| #5 | Terramate | Berlin, Germany | CLI free and open source / Cloud free tier, paid on request | EU (Germany) |
| #6 | Canonical Juju | London, United Kingdom | Free and open source; Ubuntu Pro support available | United Kingdom (adequacy decision, outside the EEA) |
| #7 | Digger | London, United Kingdom | Free and open source / Pro and Enterprise on request | United Kingdom (adequacy decision, outside the EEA) |
How each alternative compares to Terraform
- Which law reaches it. Vendor-neutral; no single jurisdiction. Terraform is run from the United States, so the CLOUD Act obliges the provider to hand over data on a valid order regardless of which country the servers are in.
- Where the data sits. Runs wherever you run it.
- Source code. Open source, where Terraform is not: you can read what it does rather than take the description on trust.
Best for: Anyone whose objection is that one company can change the licence again
OpenTofu is the fork that followed the August 2023 relicensing, now stewarded by the Linux Foundation under MPL 2.0. It reads the same HCL and the same state file, which is why this is the least painful replacement on this entire site: you install a different binary, point it at the estate you already have, and the plan output looks like the plan output you are used to.
It has also shipped things Terraform has not. State encryption is the one that matters most in practice, because Terraform state holds secrets in plain form and everybody knows it and almost nobody fixes it. Provider and module coverage runs into the thousands, and there is no telemetry and no vendor in the execution path.
Be clear about what it is not. OpenTofu has no jurisdiction, no European establishment and no company to sign a contract with, so it answers licence risk and answers nothing about supplier sovereignty. It is also only an engine — where runs happen, who approves them and where state lives are all still your decisions.
What OpenTofu does better than Terraform
- MPL 2.0 under Linux Foundation stewardship, so no single company can narrow the terms the way HashiCorp did in 2023
- Accepts your existing HCL and state file, so the switch is a binary change rather than a rewrite
- State encryption is built in, which Terraform still does not offer
- No telemetry and no licensor whose commercial interests sit next to your infrastructure code
- Free, with no per-resource meter of the kind HCP Terraform bills against
Where OpenTofu is a step down from Terraform
- Not European in any sense: no jurisdiction, no establishment, no contracting party for procurement
- The registry ecosystem still centres on HashiCorp, and some commercial providers were slow to certify OpenTofu
- An engine only, so remote runs, approvals and state storage need a separate answer
- Divergence from Terraform grows with every release, so long-term portability between the two is decreasing
Standout against Terraform. It is the only replacement here that requires no migration at all, because it was forked from the thing you are leaving and never changed the file format.
- Which law reaches it. EU (Poland) engineering, US-registered entity. Terraform is run from the United States, so the CLOUD Act obliges the provider to hand over data on a valid order regardless of which country the servers are in.
- Where the data sits. EU region available; self-hosted workers keep credentials local.
- Source code. Closed source, as Terraform is.
Best for: Platform teams whose real dependency is Terraform Cloud, not the CLI
Spacelift is the managed platform answer: remote runs, drift detection, stack dependencies and Open Policy Agent rules that block a non-compliant plan instead of noting it. It orchestrates OpenTofu, Terraform, Pulumi, CloudFormation and Ansible, which is the practical shape of most estates that never standardised on one tool.
The feature that matters against HCP Terraform is the self-hosted worker pool: the control plane is managed while the workers run inside your own VPC, so the credentials that can create and destroy anything in your cloud account never leave your network. A Self-Hosted edition goes further, and is available on the Enterprise+ tier only.
The pricing shape is the honest warning. There is a free plan with two users and one public worker, and the next published step is Starter+ at $20,000 a year, with Business and Enterprise quoted on request. Against HCP Terraform's per-resource meter that flips the arithmetic completely: a large estate may pay less here and a small team will pay a great deal more. Spacelift Inc. is also a US-registered entity, with its engineering in Warsaw.
What Spacelift does better than Terraform
- Self-hosted workers keep cloud credentials inside your own VPC, which HCP Terraform's managed runs do not
- OPA policies that actually block a plan, rather than Sentinel under a HashiCorp licence
- Orchestrates Pulumi, CloudFormation and Ansible alongside OpenTofu, where HCP Terraform is single-engine
- European engineering and operations in Warsaw with an EU region available
- Priced by contract rather than by resources under management, which suits a large estate with a flat budget
Where Spacelift is a step down from Terraform
- Spacelift Inc. is US-registered, so an EU contracting party is not what you are getting
- The published jump from the free plan to $20,000 a year is steep for a team of five
- Not open source, so you are exchanging one proprietary platform for another
- Self-hosting is confined to the top tier, unlike the tools here that simply run in your CI
Standout against Terraform. It is the only managed platform here that gives you the policy engine and the approvals without ever handing it the keys to your cloud account.
- Which law reaches it. EU (Netherlands). Terraform is run from the United States, so the CLOUD Act obliges the provider to hand over data on a valid order regardless of which country the servers are in.
- Where the data sits. Your own infrastructure.
- Source code. Open source, where Terraform is not: you can read what it does rather than take the description on trust.
Best for: Teams whose Terraform is mostly configuring machines rather than creating them
NixOS does not provision cloud resources and this page is not going to claim it does. It belongs here because a large share of what people actually keep in Terraform is machine configuration — packages, services, users, firewall rules — expressed through provisioners and templates because the tool that created the VM was already open.
For that job NixOS is structurally better. The whole system is declared in configuration, each change builds a new generation alongside the previous one, and switching is atomic, so rolling back is certain rather than hopeful. Nix development shells give byte-identical dependencies on every machine, which removes an entire class of problem instead of containerising around it.
It is stewarded by the NixOS Foundation in Utrecht under an MIT licence, with no commercial vendor anywhere in the chain — the cleanest licence position on this page. The costs are real: a genuinely steep learning curve, an unusual language, uneven documentation, and a filesystem layout that breaks software expecting ordinary paths.
What NixOS does better than Terraform
- MIT-licensed under a Dutch foundation, with no licensor able to change the terms the way IBM now could
- Rollback is guaranteed because the previous system generation was never overwritten, which no Terraform workflow gives you
- Reproducible development environments, which Terraform has never attempted
- Runs entirely on your own infrastructure with no vendor, meter or account involved
Where NixOS is a step down from Terraform
- Does not provision cloud resources, so it replaces part of your Terraform usage and not the tool
- A steep learning curve and a language nobody already knows, where HCL is close to universal
- Documentation is uneven and the non-standard paths break third-party software
- The investment only repays itself at a certain scale, below which it is overhead
Standout against Terraform. It is the only tool on this page where undoing a change is a guarantee rather than a second apply that you hope reverses the first.
- Which law reaches it. EU (Poland). Terraform is run from the United States, so the CLOUD Act obliges the provider to hand over data on a valid order regardless of which country the servers are in.
- Where the data sits. Cloud, or your own infrastructure and region via BYOC.
- Source code. Closed source, as Terraform is.
- Independently checked. GDPR.
Best for: Teams leaving HCP Terraform who need a pipeline more than a policy engine
Leaving HCP Terraform leaves a hole with a simple shape: something has to run the plan and the apply on a schedule, with credentials, on a machine. Buddy Works is a Polish CI/CD platform built for teams without a dedicated platform engineer, with cache-optimised pipelines, Docker-native execution and more than a hundred ready actions covering the common clouds and registries.
The relevant feature against a hosted Terraform product is Bring Your Own Cloud, which runs the pipelines inside your own infrastructure while Buddy operates the system. That keeps the credentials and the execution on your side of the line without making your team responsible for maintaining a build server.
It is a general CI system rather than an infrastructure-as-code platform, so there is no drift detection, no policy engine and no state management — those remain your problem or another tool's. It is not open source, pricing is by plan with BYOC on top, and the trial runs for fourteen days without a card.
What Buddy Works does better than Terraform
- Polish company under EU jurisdiction, where HCP Terraform's European option is quoted rather than listed
- Bring Your Own Cloud runs the pipeline in your infrastructure while somebody else operates the platform
- Builds Linux, Windows and macOS including ARM, which a Terraform-only platform does not cover
- A working pipeline is assembled from ready actions rather than authored as YAML by a specialist
- Handles everything else your team builds, not only the infrastructure runs
Where Buddy Works is a step down from Terraform
- Not an infrastructure-as-code platform: no state management, no drift detection, no policy enforcement
- Not open source, so the licence argument against Terraform applies here in a milder form
- A smaller ecosystem than the CI systems most teams already run
- Priced by plan rather than usage, which suits steady volume and punishes bursts
Standout against Terraform. It is the option here that solves the unglamorous half of leaving HCP Terraform: somewhere for the apply to run that you do not have to operate yourself.
- Which law reaches it. EU (Germany). Terraform is run from the United States, so the CLOUD Act obliges the provider to hand over data on a valid order regardless of which country the servers are in.
- Where the data sits. EU; the CLI runs in your own CI.
- Source code. Open source, where Terraform is not: you can read what it does rather than take the description on trust.
Best for: Estates where the same provider block has been copied into a hundred directories
Terramate GmbH is in Berlin, and it is the only supplier on this page that is both EU-established and directly useful whether you stay on Terraform or move to OpenTofu. If your requirement is a European contracting party rather than a foundation licence, this is the entry to read twice.
The problem it solves appears at scale. In a large estate the backend configuration, provider block and tagging convention are repeated in every stack, and changing them means editing hundreds of files.
Terramate generates that from definitions, and the output is ordinary readable Terraform rather than an abstraction, so a reviewer still sees exactly what will run. Change detection then works out which stacks a commit touches and runs only those, which is the difference between a fifty-minute pipeline and a five-minute one.
The CLI is MPL 2.0 open source and runs inside the CI you already have, so no new credentials are issued to anybody. Terramate Cloud is optional. It is unnecessary for a small estate, it adds a generation step the team has to learn, and the Cloud side is less mature than Spacelift's.
What Terramate does better than Terraform
- A German GmbH, so there is an EU-established company to contract with, which neither OpenTofu nor Spacelift offers
- MPL 2.0 CLI, so the licence risk that prompted this page does not exist here either
- Runs in your existing CI, so no vendor is issued credentials to your cloud accounts
- Removes boilerplate across many stacks while leaving plain Terraform for reviewers to read
- Change detection makes large-estate pipelines several times faster, which HCP Terraform does not address
Where Terramate is a step down from Terraform
- Only useful alongside Terraform or OpenTofu, so it is a layer rather than a replacement
- Pointless overhead for an estate of a handful of stacks
- The code generation step is a new concept the whole team has to understand
- Terramate Cloud is younger and thinner than the managed platforms it sits beside
Standout against Terraform. It is the only entry here that satisfies both requirements at once: an open licence nobody can revoke and a European company whose name fits on a contract.
- Which law reaches it. United Kingdom (adequacy decision, outside the EEA). Terraform is run from the United States, so the CLOUD Act obliges the provider to hand over data on a valid order regardless of which country the servers are in.
- Where the data sits. Your own infrastructure.
- Source code. Open source, where Terraform is not: you can read what it does rather than take the description on trust.
Best for: Teams whose pain is operating what Terraform created, not creating it
Juju starts where Terraform stops. Terraform creates the database; Juju encodes how to upgrade it, back it up, scale it and recover it, as charmed operators rather than as a runbook nobody opens. For a complex multi-component application that operational knowledge is the expensive part, and Terraform has never tried to hold it.
Its other idea is relations: connect an application to a database and Juju exchanges the credentials and connection details itself, which is the glue code that otherwise sits in a deployment script slowly going stale. One model covers public cloud, Kubernetes and bare metal from a single controller, which suits a mixed estate.
Canonical Ltd is a London company, so this is the UK adequacy decision rather than EU establishment. The licence is AGPLv3, which is a genuinely open answer to the BSL question, and commercial support is available through Ubuntu Pro. The concepts are unfamiliar, the charm ecosystem is much smaller than Terraform's provider catalogue, and it is tied closely to the Canonical stack.
What Canonical Juju does better than Terraform
- AGPLv3, so the licence cannot be narrowed by a single company the way Terraform's was
- Encodes upgrades, backups and recovery as software, which Terraform leaves to documentation
- Relations wire applications together automatically instead of through hand-written glue
- One model across cloud, Kubernetes and bare metal, with commercial support available behind it
Where Canonical Juju is a step down from Terraform
- Not a provisioning replacement: it is the wrong tool for creating a VPC
- A London company under adequacy rather than an EU-established supplier
- The charm ecosystem is a fraction of the size of Terraform's provider registry
- Unfamiliar concepts with a real learning curve, and a strong pull towards the Canonical stack
Standout against Terraform. It is the only tool here that treats day-two operations as code, which is the work Terraform hands back to you the moment the apply finishes.
- Which law reaches it. United Kingdom (adequacy decision, outside the EEA). Terraform is run from the United States, so the CLOUD Act obliges the provider to hand over data on a valid order regardless of which country the servers are in.
- Where the data sits. Your own CI infrastructure.
- Source code. Open source, where Terraform is not: you can read what it does rather than take the description on trust.
Best for: Teams who want the Terraform Cloud workflow without a vendor in the blast radius
Digger gives you pull-request plan and apply, the plan posted as a comment, OPA policies, drift detection and state locking — and runs all of it inside your own GitHub Actions or GitLab CI. The credentials that can create and destroy everything you own never leave infrastructure you already secure, audit and monitor.
Set against HCP Terraform that is the whole argument. A hosted infrastructure platform necessarily holds the most powerful access in the organisation, and Digger's answer is to decline the responsibility rather than to promise it will be careful with it. It is Apache 2.0 licensed with a self-hostable orchestrator, and there is no per-run meter, so cost scales with your CI minutes rather than with a vendor's pricing model.
It is a small London company founded in 2022, which belongs in the decision honestly, and the UK adequacy decision applies rather than EU establishment. The feature set is narrower than Spacelift's and it inherits whatever concurrency limits your CI already imposes.
What Digger does better than Terraform
- Cloud credentials never leave your own CI, where HCP Terraform necessarily holds them
- Apache 2.0 open source with a self-hostable orchestrator, against a BSL-licensed incumbent
- No per-resource meter, so the bill does not grow with the size of your estate
- Reuses the CI you already audit, so no new system enters the compliance scope
- Gives you the plan-on-pull-request workflow people actually stay on HCP Terraform for
Where Digger is a step down from Terraform
- A narrower feature set than Spacelift, with less depth in policy and dependency handling
- Inherits your CI's concurrency limits, so a busy estate queues behind everything else you build
- A small company founded in 2022, which is a genuine consideration for production tooling
- UK jurisdiction under adequacy rather than an EU-established supplier
Standout against Terraform. It is the only tool here that delivers the hosted Terraform experience while ensuring that nobody outside your own systems ever holds the keys.
What actually breaks when you switch
The engine swap is the easy part and people over-plan it. OpenTofu takes your HCL and your state unchanged, so the actual work is auditing pinned providers, updating CI images and checking any commercially supported module whose vendor certified Terraform by name. A small estate is a day. What catches teams out is a feature added to one project after the fork that the other never gained.
Leaving HCP Terraform is the harder move, and it is the one most people are actually making. State has to be relocated to a backend you control and encrypt, run pipelines rebuilt somewhere, Sentinel policies re-expressed in Open Policy Agent, and the private module registry replaced. None of that is difficult in isolation and all of it lands on the same team in the same fortnight.
The quiet cost is the registry. Providers and modules are published to HashiCorp's registry first, and some vendors were slow to state OpenTofu support explicitly even where the code works. Take the inventory before you commit, because the module with a support contract attached is the one that stops the migration two weeks in.
Is Terraform still open source?
No, and this is the most common thing people get wrong about it. Terraform was MPL 2.0 until 10 August 2023 and has been under the Business Source License 1.1 since. The BSL is source-available: you can read the code and use it within the granted terms, and each published version converts to MPL 2.0 four years after it ships.
The distinction matters because the two licences answer different questions. An open-source licence tells you what you may do and cannot be narrowed later for code already released. A source-available licence with a delayed conversion tells you what the licensor currently permits, and the licensor is now IBM.
If your organisation has an open-source policy that new dependencies must satisfy, Terraform no longer satisfies it and OpenTofu does. Several procurement teams discovered this only when the policy check ran.
Does the BSL actually stop my company using Terraform?
Almost certainly not, and it would be dishonest to imply otherwise. The additional use grant permits production use; what it excludes is offering the licensed work to third parties on a hosted or embedded basis in competition with IBM's commercial products. A platform team provisioning its own company's infrastructure is not doing that.
The teams that genuinely need to look are the ones selling something built on top: managed platform vendors, consultancies offering a hosted run service, anyone embedding the binary in a product customers buy. That group needed the October 2023 clarifications and still reads the grant carefully.
For everyone else the argument is about risk rather than compliance. The terms changed once, without notice, on code that described production systems. The question is not whether today's grant covers you but who decides next year's, and the answer to that is a single licensor rather than a foundation.
Do I have to rewrite my infrastructure code to leave?
No, and this page is unusual on this site for saying so. OpenTofu was forked from Terraform and accepts the same HCL and the same state file, so for an estate written against Terraform 1.5 the move is installing a different binary and running it against what you already have. There is no data migration and no rewrite.
Two things do need checking. Terraform and OpenTofu have both shipped features since the fork, so code that uses something added to Terraform after 1.5 may not have an OpenTofu equivalent and vice versa — OpenTofu, for instance, added state encryption, which Terraform does not have. And the provider and module ecosystem still orbits HashiCorp's registry, so verify that every provider you pin is resolvable from where OpenTofu fetches it.
Budget a day for a small estate and a sprint for a large one, and spend most of that time on CI pipelines and the provider inventory rather than on the code.
If OpenTofu is not European, why is it first on this page?
Because on this page the thing being escaped is a licence controlled by one company, and the fix for that is governance rather than geography. OpenTofu is stewarded by the Linux Foundation under MPL 2.0 with many contributing organisations, which means no single participant can relicense it, because no single participant owns it. That is a structurally different guarantee from picking a vendor in a friendlier country.
It is worth being precise about what that does not give you. OpenTofu has no jurisdiction of its own, no European establishment, and no contracting party to put on a procurement form. If your requirement is an EU-established supplier with an EU contract, OpenTofu satisfies none of it and Terramate GmbH in Berlin or Buddy in Poland do.
The honest reading is that this category has two separate requirements that people conflate. Licence risk is answered by foundation governance. Supplier jurisdiction is answered by a European company. Only Terramate answers both at once, and it is an orchestration layer rather than an engine.
Which one to pick
If your objection is that a single company can change the terms of the code describing your production estate, OpenTofu is the complete answer and the move costs you a day. Foundation governance rather than a friendlier country is what actually fixes that problem.
If your procurement rule requires a European supplier with a European contract, OpenTofu does not satisfy it and Terramate GmbH does, while leaving the engine decision open. That is the entry to look at when the requirement comes from legal rather than from engineering.
If what you are really leaving is HCP Terraform, decide first who should hold your cloud credentials. Spacelift gives you the most capable platform with workers inside your own VPC; Digger gives you the same workflow with nothing held by anyone; Buddy Works gives you somewhere to run it when the policy engine was never the point.
And if you are staying, stay deliberately. Read the additional use grant against what your company actually sells, note that the Licensor is now IBM, and record the decision — because the value of having read it is that the next licence change finds you already knowing what it changes.
Frequently Asked Questions
OpenTofu if you are replacing the engine, because it takes your existing code and state unchanged. Spacelift if you are replacing HCP Terraform and want policy, approvals and drift detection as a managed platform. Digger if you want that workflow without any vendor holding cloud credentials. Terramate if your problem is that a hundred stacks contain the same copied boilerplate. NixOS and Juju are not Terraform replacements at all and are on the page for a different job.
On 10 August 2023, when HashiCorp relicensed its products from MPL 2.0 to the Business Source License 1.1, with clarifications published on 16 October 2023. The LICENSE file in the Terraform repository now names International Business Machines Corporation as Licensor, following IBM's completion of the acquisition on 27 February 2025.
Yes, with OpenTofu, which reads the same state format. That is what makes this the easiest migration on this site. What you should think about separately is where that file lives, because state contains an inventory of everything you own and secrets in plain form. If it currently sits in HCP Terraform, moving it back to your own encrypted backend is the more consequential half of the change.
Mostly, and the exceptions are worth finding early. OpenTofu resolves the same provider protocol and carries thousands of providers and modules, but the ecosystem still gravitates to HashiCorp's registry and some commercial vendors were slow to certify OpenTofu explicitly. Inventory your pinned providers before you switch, and check any vendor-supplied module that comes with a support contract attached.
It is billed per resource under management: Essentials from $0.10 per resource per month, Standard from $0.47 and Premium from $0.99, with Enterprise for self-managed deployments priced on request. New accounts get a $500 credit across the IBM HashiCorp Cloud Platform. The shape matters more than the rate — your bill grows with the size of your estate rather than with the size of your team.
There is an IBM HCP Terraform Europe offering for enterprises that need geographic service locality, and it carries custom pricing rather than a published rate. So European locality exists and you cannot select it yourself; you acquire it through a sales process. If your requirement came from a procurement rule, that route is open. If it came from wanting to avoid a US-controlled supplier, buying a European region from IBM does not address it.
Less than it used to be, and the risk is a different one from what people assume. OpenTofu is not a weekend fork: it sits under Linux Foundation stewardship with contributions from many companies, which is precisely the structure that makes a repeat relicensing impossible. The real risk is divergence over time, as each project adds features the other does not have, so code written for one gradually stops being portable to the other.
Three things worth naming. The gravity of the registry, where providers and modules are published first. Commercial support with IBM behind it, which some enterprise procurement processes require by name. And the surrounding products — Vault, Consul, Packer — which are under the same licence and which nothing on this page replaces. If your platform depends on all three, this is a programme rather than a switch.
For the CLI, a day to a sprint, because the code and state come across unchanged and the work is in CI pipelines and the provider inventory. For HCP Terraform, plan in weeks: state has to be relocated, run pipelines rebuilt, policies re-expressed in whatever engine the new platform uses, and the private module registry replaced. The engine is easy and the platform around it is not, which is the reverse of what most people expect.
Who worked on this review
Explore More European Alternatives
Discover privacy-focused European alternatives to other popular US tech services.