Mr.PlanB Logo
    Infrastructure Automation Ledger 2026
    Ledger
    Infrastructure Automation Ledger 2026

    OpenTofu: Infrastructure Automation Ledger 2026 Scorecard

    OpenTofu is the MPL 2.0 continuation of Terraform 1.5 with published support windows, state encryption and no vendor; orchestration lives in other products.

    Edition October 16, 20268 min read

    Where OpenTofu put its effort, 2025 to 2026

    Each square is one capability area in one year. The brighter it is, the more the area gained in the Infrastructure Automation Ledger that year; 1, 2 and 3 mark that year's biggest pushes.

    ’25’26
    Provisioning breadth2
    Configuration management depth
    State, planning and safety
    Language and authoring
    Orchestration and self-service
    Scale and control plane
    Security and compliance3
    Bare metal and day-2 operations
    Ecosystem and integrations
    AI assistance1
    no change+0.5+1+1.5 or moreScores out of 5 per area.

    Biggest pushes overall: AI assistance (+1.5), Provisioning breadth (+0.5), Security and compliance (+0.5). Already strong in 2024: Language and authoring, State, planning and safety, Provisioning breadth.

    Top focus each year: 2025 AI assistance. 1 of 2 years show no gain in any area.

    Part of the Infrastructure Automation Ledger (edition 2026-10-16). How the scores work: methodology.

    Snapshot

    • Type: community project (foundation hosted, no first-party commercial product)
    • Owner: OpenTofu, a Series of LF Projects, LLC, under the Linux Foundation since 2023-09-20; CNCF Sandbox since 2025-04-23
    • Engine: the project is the engine; several IaC management platforms run it
    • Licence: MPL 2.0 since the fork from Terraform 1.5.x, no change; registry data Apache 2.0; no editions, no paid tier
    • Latest release: v1.13.1 (2026-10-01); v1.13.0 (2026-09-30); v1.12.8 (2026-10-14)
    • Support lifecycle: Each minor series supported about 12 to 13 months with published dates: 1.13 to 2027-08-01, 1.12 to 2027-02-01; two or three series supported at once; in-place minor upgrades
    • Next milestone: 1.14.0, listed as unreleased in the changelog with support stated until 2028-02-01; no release date published
    • Targets: Same upstream providers as Terraform through registry.opentofu.org: vSphere vmware/vsphere, VMware by Broadcom, v2.17.1 (2026-09-14); Proxmox VE bpg/proxmox, community, v0.116.1 (2026-10-15); Nutanix nutanix/nutanix, vendor, v2.5.0 (2026-10-07); Hyper-V taliesins/hyperv, community, last release 2024-02-11; XCP-ng vatesfr/xenorchestra, Vates organisation, v0.41.2 (2026-09-15); KubeVirt no maintained provider. Whether each vendor tests against OpenTofu was not found
    • Tier: Infrastructure-as-code engine
    • Archetype: Dependable builder
    • Evidence grade: A. Release changelogs per branch, the repository licence, the project blog, endoflife.date support windows and the CNCF project page were read as primary sources. Fork and Linux Foundation dates and the CNCF licence exception are grade B. Whether provider vendors test against OpenTofu is unverified (grade C).

    Scorecard

    DimensionScoreDirection
    Capability today49.5 / 100higher is better
    Momentum since 2024+5.4 pointshigher is better
    Credibility20.0 / 25higher is better
    Sustainability risk7.5 / 25lower is better
    Technical lock-in4.0 / 25lower is better
    Integrator fit9.0 / 20higher is better
    AI leverage2.5 / 20higher is better
    Ledger Index55.8 / 100balanced weights; see the methodology

    Capability by domain, 2015 to today

    Scores 0 to 5 against fixed anchors; total is weighted to 100.

    Domain (weight)201520192022Today
    Provisioning breadth (15)n/an/an/a4
    Configuration management depth (12)n/an/an/a0.5
    State, planning and safety (12)n/an/an/a4
    Language and authoring (10)n/an/an/a4.5
    Orchestration and self-service (10)n/an/an/a0.5
    Scale and control plane (8)n/an/an/a2
    Security and compliance (10)n/an/an/a2.5
    Bare metal and day-2 operations (8)n/an/an/a0.5
    Ecosystem and integrations (8)n/an/an/a3.5
    AI assistance (7)n/an/an/a1.5
    Total / 100n/an/an/a49.5

    Year by year: 2024: 44.1 · 2025: 49.5 · 2026: 49.5

    What moved the score

    YearDomainChangePointsTrigger
    2025AI assistance0 → 1.5+2.1Official registry MCP server 1.0.0 (2025-06-06): documentation lookup only, no plan explanation, identity or audit
    2025Ecosystem and integrations3 → 3.5+0.8CNCF Sandbox (2025-04-23); 1.10: OCI module sources and provider mirrors
    2025Provisioning breadth3.5 → 4+1.5vSphere provider maintained by VMware by Broadcom (2025-05-14) and Nutanix vendor provider serve both engines
    2025Security and compliance2 → 2.5+1.01.11: ephemeral values and resources, write-only attributes

    Direction, 2023 to 2026

    Each release item carries one theme and one driver (V vision, C customer need, M market window, U upstream, P portfolio).

    DateItemThemeDriver
    2023-08Fork of Terraform announced in response to the BUSL changeLicensingM
    2023-09Joins the Linux Foundation as OpenTofuPackagingM
    2024-011.6.0 GA on the announced date, with its own registry and test frameworkPackagingM
    2024-041.7: end to end state encryption, removed block, provider-defined functionsSecurityC
    2024-04Public response to HashiCorp's cease and desist letter over the removed blockLicensingM
    2024-071.8: test mocking and overrides, variables in module sources, .tofu filesLanguageC
    2024-09Building the OpenTofu Registry on Cloudflare R2EcosystemC
    2025-011.9: for_each on provider configuration, -exclude optionProvisioning breadthC
    2025-04Accepted into the CNCF at Sandbox levelEcosystemV
    2025-06OpenTofu MCP server 1.0.0 for registry lookupAIV
    2025-061.10: OCI module sources and provider mirrors, S3 locking without DynamoDB, deprecated variablesEcosystemC
    2025-121.11: ephemeral values and resources, write-only attributes, enabled meta-argumentSecurityU
    2026-01Machine and human readable output streamsOrchestrationC
    2026-051.12: destroy lifecycle argument, const variables, identity-based import, workload identity for backendsState and planningC
    2026-091.13: experimental linting and symbol libraries, convert function, crash recovery state fileLanguageV
    • Centre of gravity: Ecosystem (3), Licensing (2), Packaging (2)
    • Blind spots (no items): Configuration management, Scale, Bare metal
    • Driver mix: Vision 20%, Customer need 47%, Market window 27%, Upstream 7%, Portfolio 0%
    • Idea-to-need index: 0.21 (0 = demand-pulled, 1 = thesis-pushed)

    Credibility: 20.0 / 25

    CriterionScoreBasis
    Cadence4A minor every five to six months (2025-01-09, 2025-06-24, 2025-12-09, 2026-05-14, 2026-09-30) with betas and release candidates; support dates are published, release dates are not
    Roadmap transparency4Open repository, RFC process and issue tracker, project blog; no dated roadmap file
    Say-do41.6.0 GA held its announced date; state encryption, ephemeral values and OCI sources shipped; linting and symbol libraries remain experimental
    Velocity4Several substantive features in each minor since 1.7
    Lifecycle stability4About 12 months of support per series with published end dates and two or three series supported at once; the 1.11 end date differs by 19 days between the changelog and endoflife.date

    Sustainability risk: 7.5 / 25 (lower is better)

    CriterionScoreBasis
    Maintainer concentration2.5Technical Steering Committee membership and funders not found on the project site (the about page returned 404); governance sits with LF Projects, LLC
    Backer health1.5Linux Foundation project since 2023-09-20 and CNCF Sandbox since 2025-04-23; no sponsor list read
    Licence-change risk1MPL 2.0 held by a foundation; the project exists because of a licence change and its manifesto commits to staying open source
    Abandonment signs0.5Five minors in 21 months, patch releases on two series in October 2026, support windows published to 2028
    Security response2No advisory process read in this pass; patches land on supported series (1.12.8 on 2026-10-14) and 1.13 added crash recovery to errored.tfstate

    Technical lock-in: 4.0 / 25 (lower is better)

    CriterionScoreBasis
    Formats1HCL and JSON state shared with Terraform; a state encrypted by OpenTofu is readable only by OpenTofu
    Export path1.5Back to Terraform only before diverging features (enabled, state encryption, provider for_each) are used; Pulumi converts HCL
    Stack coupling0.5Any backend, any provider
    Hardware / cloud coupling0Fully offline capable; no cloud service
    Skills and tooling coupling1Terraform skills transfer directly

    Integrator fit: 9.0 / 20

    CriterionScoreBasis
    Partner programme openness3No programme and no vendor; MPL 2.0 carries no competitive-use clause, so hosted platforms can run and resell it
    Multi-tenancy and self-service0.5Engine only; runs, approvals and tenancy come from third-party platforms
    API and infrastructure-as-code3CLI with JSON output and machine readable streams; no API server
    Skills and certification2.5No OpenTofu exam found; Terraform certifications and the Terraform talent pool apply

    AI leverage: 2.5 / 20

    How the product uses AI, scored per year like capability. Only GA counts.

    Criterion20192022TodayLatest step
    AI code and plan generationn/an/a02024: None
    AI platform stackn/an/a02024: None
    AI-assisted operationsn/an/a02024: None
    Agent opennessn/an/a2.52025: Official MCP server 1.0.0 (2025-06-06) reads registry data and documentation through five tools; no identity or audit layer; no release since

    Year by year: 2024: 1 · 2025: 2.5 · 2026: 2.5

    AI say-do, 2023 to 2026: 1 AI items announced; 1 GA, 0 still preview, 0 pending, 0 dropped. The one AI item is a documentation lookup server; whether it is still maintained is an open question.

    OpenTofu has no AI inside the engine and states no plan for one. Its MCP server answers registry and documentation questions for an assistant and does not touch infrastructure.

    Reading

    OpenTofu matches Terraform's CLI on breadth, language and state, and in two places goes further: state and plan encryption since 1.7 and published support windows for every series. It scores low on orchestration, scale and security because those live in the platforms that run it, which is the point of an engine with no vendor. The sustainability record is good on cadence and delivery and thin on governance disclosure: who sits on the steering committee and who funds the work could not be read from the project site. Check which features you use before assuming a return path to Terraform, because encrypted state and the enabled meta-argument do not travel.

    Open questions and items to re-score

    • Technical Steering Committee membership and project funding (about page 404).
    • Whether vSphere, Nutanix and other provider vendors test or support OpenTofu explicitly.
    • Whether the MCP server is maintained (no release after v1.0.0 on 2025-06-06).
    • The exact fork point (Terraform 1.5.6 or another 1.5.x).
    • The 1.11 support end date: 2026-08-01 in the changelog, 2026-08-19 on endoflife.date.
    • Linting and symbol libraries are experimental in 1.13; re-score the language domain when they reach GA.

    Related reading

    More from Infrastructure Automation Ledger 2026