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

    Tinkerbell: Infrastructure Automation Ledger 2026 Scorecard

    Tinkerbell is the CNCF Sandbox bare-metal workflow engine that outlived its sponsor Equinix Metal; active 0.x releases, no plan mode, no vendor.

    Edition October 16, 20267 min read

    Where Tinkerbell put its effort, 2021 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.

    ’21’22’23’24’25’26
    Provisioning breadth
    Configuration management depth
    State, planning and safety
    Language and authoring1
    Orchestration and self-service2
    Scale and control plane3
    Security and compliance
    Bare metal and day-2 operations
    Ecosystem and integrations
    AI assistance
    no change+0.5+1+1.5 or moreScores out of 5 per area.

    Biggest pushes overall: Language and authoring (+0.5), Orchestration and self-service (+0.5), Scale and control plane (+0.5). Already strong in 2020: Bare metal and day-2 operations, Ecosystem and integrations, State, planning and safety.

    Top focus each year: 2026 Language and authoring. 5 of 6 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 (CNCF Sandbox)
    • Owner: CNCF Sandbox project (accepted 2020-11-10) governed by its maintainers; origin sponsor Equinix Metal ended service on 2026-06-30; maintainer employers not found
    • Licence: Apache-2.0; free; no paid support or commercial edition found
    • Latest release: 0.25.0 (2026-07-31)
    • Support lifecycle: None published; the latest release is the one maintained; all releases are 0.x
    • Next milestone: none published (a roadmap repository exists, not read)
    • Targets: Bare metal only, through network and ISO boot with BMC control over Redfish, IPMI and Intel AMT (Raspberry Pi netboot since 0.25); Cluster API provider 0.7.0 (2026-04-27); no VM, cloud, vSphere, Proxmox VE or Nutanix integration
    • Tier: Host and bare-metal provisioning
    • Archetype: Slow follower
    • Evidence grade: B. Release feeds, the governance document, the CNCF project page and the two 2026 security advisories were read from primary sources. Feature descriptions come from the repository README because the documentation site returned no content. Releases from 2021 to 2024 were not researched, so capability before 2026 is held flat on inferred values.

    Scope of the scores: Tinkerbell has no plan or dry-run mode and no drift detection of installed hosts. Desired state lives in Hardware, Template and Workflow objects that controllers reconcile, which is the reading of inclusion test 4 applied to the bare-metal family. It needs Kubernetes. Scored with those limits.

    Scorecard

    DimensionScoreDirection
    Capability today29.7 / 100higher is better
    Momentum since end of 2022+3.6 points (2020 to 2022: +0.0)higher is better
    Credibility12.0 / 25higher is better
    Sustainability risk11.0 / 25lower is better
    Technical lock-in9.5 / 25lower is better
    Integrator fit6.5 / 20higher is better
    AI leverage2.0 / 20higher is better
    Ledger Index36.9 / 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/a0.50.5
    Configuration management depth (12)n/an/a11
    State, planning and safety (12)n/an/a1.51.5
    Language and authoring (10)n/an/a1.52
    Orchestration and self-service (10)n/an/a1.52
    Scale and control plane (8)n/an/a11.5
    Security and compliance (10)n/an/a1.51.5
    Bare metal and day-2 operations (8)n/an/a33.5
    Ecosystem and integrations (8)n/an/a22
    AI assistance (7)n/an/a00
    Total / 100n/an/a26.129.7

    Year by year: 2020: 26.1 · 2021: 26.1 · 2022: 26.1 · 2023: 26.1 · 2024: 26.1 · 2025: 26.1 · 2026: 29.7

    What moved the score

    YearDomainChangePointsTrigger
    2026Bare metal and day-2 operations3 → 3.5+0.80.24.0 and 0.25.0: HomeAssistant power provider, PXELinux and Raspberry Pi boot routes, Kubernetes executor for the agent
    2026Language and authoring1.5 → 2+1.00.23.0: toYaml and fromYaml template functions; Hardware attributes
    2026Orchestration and self-service1.5 → 2+1.00.23.0: Tinkerbell UI; single HTTP server for Smee and Tootles
    2026Scale and control plane1 → 1.5+0.80.23.0: configurable MaxConcurrentReconciles for Rufio controllers

    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
    2024-11Equinix tells Metal customers the service ends on 2026-06-30 (Tinkerbell's origin sponsor)EcosystemP
    2025-06CNCF community bare-metal cluster on Equinix Metal to be decommissioned by 2025-12-31EcosystemP
    2025-120.20.0 to 0.22.0: three minor releases between August and December 2025 (contents not read)PackagingC
    2026-030.23.0: single HTTP server for Smee, Tootles and the new Tinkerbell UI (breaking change)OrchestrationC
    2026-030.23.0: configurable MaxConcurrentReconciles for Rufio; Hardware attributes; toYaml and fromYaml template functionsScaleC
    2026-04cluster-api-provider-tinkerbell 0.7.0EcosystemU
    2026-06Equinix Metal service ends; resources removed 2026-07-01EcosystemP
    2026-070.24.0: fix for a critical unauthenticated SSRF in the Smee ISO handler (GHSA-f2j9-ffq8-pg75)SecurityC
    2026-070.24.0: bmclib HomeAssistant power providerBare metalV
    2026-070.25.0: fix for workflow template env exposure (GHSA-p37j-ccmf-67x4); hermetic template functionsSecurityC
    2026-070.25.0: TFTP route abstraction with PXELinux and Raspberry Pi netboot; Tink Agent per-action namespaces and Kubernetes executorBare metalC
    • Centre of gravity: Ecosystem (4), Security (2), Bare metal (2)
    • Blind spots (no items): Provisioning breadth, Configuration management, State and planning, Language, AI, Licensing
    • Driver mix: Vision 9%, Customer need 55%, Market window 0%, Upstream 9%, Portfolio 27%
    • Idea-to-need index: 0.14 (0 = demand-pulled, 1 = thesis-pushed)

    Credibility: 12.0 / 25

    CriterionScoreBasis
    Cadence2.5Irregular: four months between 0.23.0 and 0.24.0, then 0.25.0 two days later; no schedule
    Roadmap transparency3Public roadmap repository, governance document and published advisories; no dated roadmap
    Say-do2.5Not enough announcements to measure
    Velocity2.5Consolidation into one binary and a UI in 2026; otherwise fixes
    Lifecycle stability1.50.x versions, breaking changes between minors, no support windows published

    Sustainability risk: 11.0 / 25 (lower is better)

    CriterionScoreBasis
    Maintainer concentration3.5Origin sponsor Equinix Metal ended service on 2026-06-30; maintainer employers after that were not found
    Backer health3.5No vendor, no paid support; CNCF Sandbox since 2020 without incubation
    Licence-change risk0.5Apache-2.0 under the CNCF
    Abandonment signs2Three releases in 2026 show activity, but all are 0.x on an irregular rhythm
    Security response1.5A critical SSRF and a high template exposure were fixed in July 2026 within days of each other, with published GHSA advisories

    Technical lock-in: 9.5 / 25 (lower is better)

    CriterionScoreBasis
    Formats1.5Kubernetes YAML with Tinkerbell-specific workflow templates
    Export path2.5No export to other tools
    Stack coupling2.5Requires Kubernetes
    Hardware / cloud coupling0.5Any BMC with Redfish, IPMI or Intel AMT; offline capable
    Skills and tooling coupling2.5Kubernetes plus project concepts and Go templating; small community

    Integrator fit: 6.5 / 20

    CriterionScoreBasis
    Partner programme openness1None
    Multi-tenancy and self-service1.5Kubernetes namespaces and RBAC only
    API and infrastructure-as-code3Kubernetes API and Cluster API provider 0.7.0
    Skills and certification1None found

    AI leverage: 2.0 / 20

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

    Criterion20192022TodayLatest step
    AI code and plan generationn/a002020: None
    AI platform stackn/a002020: None
    AI-assisted operationsn/a002020: None
    Agent opennessn/a222020: Kubernetes API and custom resources; no MCP server found

    Year by year: 2020: 2 · 2021: 2 · 2022: 2 · 2023: 2 · 2024: 2 · 2025: 2 · 2026: 2

    AI say-do, 2023 to 2026: 0 AI items announced; 0 GA, 0 still preview, 0 pending, 0 dropped. No AI items announced.

    Tinkerbell has no AI features and no official agent interface beyond the Kubernetes API. Nothing on the record points to AI work.

    Reading

    Tinkerbell is the lightest tool in this group: a bare-metal workflow engine with network boot, BMC control and a metadata service, and little else the Ledger measures. Sustainability matters more than capability here: its origin sponsor Equinix Metal shut down on 2026-06-30, the project is still 0.x with no support windows, and the CNCF has not moved it past Sandbox in six years. On the positive side, two serious 2026 advisories were fixed within days and three releases shipped this year. Before adopting it, find out who the current maintainers work for, because the public record does not say.

    Open questions and items to re-score

    • Maintainer employers after the Equinix Metal shutdown were not found.
    • Whether the project will reach CNCF incubation; no date or application found.
    • Documentation content was not read (the docs site returned only a title); features come from the README.
    • Releases 0.20.0 to 0.22.1 (2025 to January 2026) were not read.
    • A non-Kubernetes (Docker) deployment mode was not confirmed.
    • No MCP server search was run beyond the repository.

    Related reading

    More from Infrastructure Automation Ledger 2026