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

    Pulumi: Infrastructure Automation Ledger 2026 Scorecard

    Pulumi pairs a stable Apache 2.0 engine in seven languages with a fast-moving cloud service whose tiers, prices and AI agent changed several times in 2026.

    Edition October 16, 20268 min read

    Where Pulumi put its effort, 2019 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.

    ’19’20’21’22’23’24’25’26
    Provisioning breadth3
    Configuration management depth
    State, planning and safety2
    Language and authoring
    Orchestration and self-service12
    Scale and control plane3
    Security and compliance21
    Bare metal and day-2 operations
    Ecosystem and integrations
    AI assistance11
    no change+0.5+1+1.5 or moreScores out of 5 per area.

    Biggest pushes overall: Security and compliance (+3), AI assistance (+3), Orchestration and self-service (+2.5). Already strong in 2018: Language and authoring, Provisioning breadth, State, planning and safety.

    Top focus each year: 2019 Orchestration and self-service; 2024 Security and compliance; 2025 AI assistance; 2026 AI assistance. 4 of 8 years show no gain in any area.

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

    Snapshot

    • Type: product (Apache 2.0 engine and CLI with Pulumi Cloud; self-hosting on the Enterprise edition only)
    • Owner: Pulumi Corporation, Seattle, private and venture backed; last round found is a $41 million Series C in 2023; no acquisition
    • Engine: Pulumi open source engine and CLI (Apache 2.0) with native and Terraform-bridged providers
    • Licence: Engine Apache 2.0 with no change found. Pulumi Cloud editions (2026-10-15): Free ($0, one user), Essentials ($40 a month, up to 500 resources), Pro ($400 a month, up to 2,000 resources), Enterprise ($2,000 a month, up to 4,750 resources); resources billed hourly above the included amount; Neo at $3 per million tokens on paid editions
    • Latest release: v3.269.0 (2026-10-15); minor releases roughly weekly
    • Support lifecycle: Single rolling 3.x line since 2021-04-19; no support policy page found (the documented URL returned 404)
    • Next milestone: none published
    • Targets: vSphere: pulumi/pulumi-vsphere, Pulumi maintained, bridged from the Broadcom-maintained Terraform provider, v4.18.2 (2026-10-13). Proxmox VE: muhlba91/pulumi-proxmoxve, community, v8.6.0 (2026-09-06). Nutanix: pierskarsenbarg/pulumi-nutanix, community, v0.18.0 (2026-10-02). Hyper-V: pulumi/pulumi-hyperv repository exists with no releases. XCP-ng: Xen Orchestra provider maintained by Vates, listed in the Pulumi Registry (release date not retrieved). KubeVirt: through pulumi-kubernetes v4.34.2 (2026-09-21) as custom resources, no dedicated package. Any Terraform provider can be bridged from the OpenTofu registry
    • Tier: Infrastructure-as-code engine
    • Archetype: Real innovator
    • Evidence grade: B. The pricing page, licence, release feed, MCP and drift documentation and the 1.0, 2.0 and 3.0 changelog entries were read from primary sources. The 3.x release history between 2021 and mid 2026 was not retrieved, so the yearly line between 2019 and 2025 holds few steps. Pricing history and funding come from a tracker and news (grade B); Neo's GA date and the support policy were not found (grade C).

    Scorecard

    DimensionScoreDirection
    Capability today60.7 / 100higher is better
    Momentum since end of 2022+13.9 points (2018 to 2022: +18.2)higher is better
    Credibility17.0 / 25higher is better
    Commercial risk11.0 / 25lower is better
    Technical lock-in8.0 / 25lower is better
    Integrator fit12.5 / 20higher is better
    AI leverage8.5 / 20higher is better
    Ledger Index61.2 / 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/a333.5
    Configuration management depth (12)n/a0.50.50.5
    State, planning and safety (12)n/a334
    Language and authoring (10)n/a4.54.54.5
    Orchestration and self-service (10)n/a2.52.53.5
    Scale and control plane (8)n/a333
    Security and compliance (10)n/a2.52.54
    Bare metal and day-2 operations (8)n/a0.50.50.5
    Ecosystem and integrations (8)n/a333.5
    AI assistance (7)n/a003
    Total / 100n/a46.846.860.7

    Year by year: 2018: 28.6 · 2019: 46.8 · 2020: 46.8 · 2021: 46.8 · 2022: 46.8 · 2023: 46.8 · 2024: 50.0 · 2025: 50.7 · 2026: 60.7

    What moved the score

    YearDomainChangePointsTrigger
    2019Ecosystem and integrations1.5 → 3+2.41.0: CI integrations, webhooks, GitHub app
    2019Language and authoring3.5 → 4.5+2.01.0: stable SDKs in TypeScript, JavaScript, Python and Go with the languages' own test frameworks
    2019Orchestration and self-service1 → 2.5+3.01.0: Pulumi Cloud with RBAC, REST API, webhooks and CI integrations
    2019Provisioning breadth2 → 3+3.01.0: stable providers for AWS, Azure, GCP, Kubernetes and DigitalOcean
    2019Scale and control plane1.5 → 3+2.41.0: Pulumi Cloud as the hosted control plane
    2019Security and compliance1 → 2.5+3.01.0: Pulumi Cloud RBAC and encrypted secrets
    2019State, planning and safety2 → 3+2.41.0: preview, checkpointed state in Pulumi Cloud, import
    2024Security and compliance2.5 → 3.5+2.0ESC secrets and configuration GA (2024-09-18)
    2024State, planning and safety3 → 3.5+1.2Insights 2.0 preview with inventory, policies and remediation (half step)
    2025AI assistance0 → 0.5+0.7Neo in public preview (2025-09-16)
    2026AI assistance0.5 → 3+3.5Neo CLI, GitHub and Slack apps (2026-05); vendor states Neo generally available; hosted MCP server with scoped tokens
    2026Ecosystem and integrations3 → 3.5+0.8Any Terraform provider from the OpenTofu registry; Terraform state and HCL support; Neo GitHub and Slack apps
    2026Orchestration and self-service2.5 → 3.5+2.0Deployments gain drift remediation and VCS settings (3.264.0), scheduled tasks and agent accounts (2026-05)
    2026Provisioning breadth3 → 3.5+1.5Terraform state, cross-language modules and HCL support (2026-08-04); community Nutanix package v0.18.0 (2026-10-02)
    2026Security and compliance3.5 → 4+1.03.258.0: opt-in credential encryption; policy violations in pull request checks (2026-09-14)
    2026State, planning and safety3.5 → 4+1.23.264.0: deployment settings for drift remediation; Terraform state support (2026-08-04)

    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
    2023-10Pulumi ESC secrets and configuration service, previewSecurityV
    2024-09ESC generally availableSecurityV
    2024-09Insights 2.0 inventory, policies and remediation, previewState and planningV
    2025-09Neo AI platform engineering agent, public previewAIV
    2026-05Neo CLI, GitHub and Slack apps, scheduled tasks, agent accounts, MCP bridgeAIV
    2026-06Neo billed per tokenPackagingP
    2026-07Migrate CloudFormation and ARM estates with discovered stacksProvisioning breadthM
    2026-08Full support for Terraform state, cross-language modules and HCLEcosystemM
    2026-08Experimental Rust support; Context APILanguageV
    2026-083.258.0: opt-in credential encryption with an OS-protected keySecurityC
    2026-08Neo Security, research preview for invited customersAIV
    2026-09Policy violations shown in pull request comments and checksSecurityC
    2026-093.264.0: deployment settings for drift remediation and VCSState and planningC
    2026-09Tiers renamed; Enterprise moves to a $2,000 a month self-serve basePackagingP
    2026-103.268.0: Automation API gains version, whoami and installOrchestrationC
    • Centre of gravity: Security (4), AI (3), State and planning (2)
    • Blind spots (no items): Configuration management, Scale, Bare metal, Licensing
    • Driver mix: Vision 47%, Customer need 27%, Market window 13%, Upstream 0%, Portfolio 13%
    • Idea-to-need index: 0.54 (0 = demand-pulled, 1 = thesis-pushed)

    Credibility: 17.0 / 25

    CriterionScoreBasis
    Cadence4A minor roughly every week (3.266.0 on 2026-09-30 to 3.269.0 on 2026-10-15); the rhythm is visible but not published
    Roadmap transparency3Open repository, changelog, blog and release notes; no dated roadmap found
    Say-do3ESC went from preview to GA in 11 months; Insights 2.0 and Neo have no dated GA announcement, so one of three trackable previews is clearly closed
    Velocity4.5Many capabilities a month across the engine, Deployments and Neo
    Lifecycle stability2.5One rolling 3.x line with no forced upgrades, but no support policy page and tiers renamed twice in 2026

    Commercial risk: 11.0 / 25 (lower is better)

    CriterionScoreBasis
    Price and licence volatility3.5Engine licence unchanged; the cloud service added per-token Neo pricing (2026-06-17), renamed every tier (2026-09-21) and repriced Enterprise (2026-09-28)
    Purchase constraints2Free edition for one user; self-hosting only on Enterprise at $2,000 a month; compliance policy packs and SCIM on Enterprise only
    Channel and access1Open source CLI, self-serve signup, AWS Marketplace listing; partner terms not published
    Owner stability2.5Independent venture-backed company; last funding round found is 2023; no acquisition
    Cost of staying supported2No version windows to manage; paid support add-ons from $625 a month (tracker, not read at Pulumi)

    Technical lock-in: 8.0 / 25 (lower is better)

    CriterionScoreBasis
    Formats2Programs in general-purpose languages; JSON checkpoint state is documented
    Export path2.5No documented export to another IaC tool; the conversion path runs into Pulumi, not out of it
    Stack coupling0.5Any backend (Pulumi Cloud, S3 and compatible, Azure Blob, GCS, PostgreSQL, local file)
    Hardware / cloud coupling1.5Pulumi Cloud is the default and hosts ESC, Deployments, Neo and Insights; self-managed backends cannot recover from some partial failures
    Skills and tooling coupling1.5Ordinary programming skills plus the Pulumi resource model

    Integrator fit: 12.5 / 20

    CriterionScoreBasis
    Partner programme openness2.5Partners page lists cloud, technology and integrator partners; no tiers, criteria or certification stated
    Multi-tenancy and self-service3.5Organisations, teams, RBAC with custom roles (25 on Pro, unlimited on Enterprise), SCIM on Enterprise, templates
    API and infrastructure-as-code4.5REST API plus the Automation API in Node.js, Python and Go
    Skills and certification2No certification found; general language skills are common, Pulumi-specific experience is rarer

    AI leverage: 8.5 / 20

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

    Criterion20192022TodayLatest step
    AI code and plan generation0032026: Neo generates and applies infrastructure changes inside Pulumi Cloud under RBAC and policy; the vendor states it is generally available, billed per token; Neo CLI since 2026-05-20
    AI platform stack0002018: None
    AI-assisted operations0022026: Neo scheduled tasks and pulumi do (2026-05) run operations work; Neo Security is a research preview (2026-08-28)
    Agent openness223.52026: Hosted MCP server at mcp.ai.pulumi.com with organisation-scoped tokens, agent accounts (2026-05); the server does not run pulumi up itself and its GA status is not stated

    Year by year: 2018: 1 · 2019: 2 · 2020: 2 · 2021: 2 · 2022: 2 · 2023: 2 · 2024: 2 · 2025: 2.5 · 2026: 8.5

    AI say-do, 2023 to 2026: 4 AI items announced; 2 GA, 2 still preview, 0 pending, 0 dropped. Neo's GA rests on the vendor product page and per-token billing; no dated GA announcement was found. The MCP bridge and Neo Security remain previews.

    Pulumi is the only product in this group with an in-product agent that writes and applies changes, and the only one that bills AI by the token. The buyer-side gap is proof: no dated GA post for Neo, no stated GA for the MCP server, and the benchmark that supports the claims is the vendor's own.

    Reading

    Pulumi's engine has not changed its licence or its major version since 2021, and the language and authoring score has been near the top of the group since 1.0. What moves is the cloud service: ESC, Insights, Deployments and Neo arrived as vision items, and in 2026 the tiers, prices and the AI agent all changed within four months. The commercial risk score sits below Terraform's because the engine is Apache 2.0 and the owner is unchanged, and above OpenTofu's because self-hosting, drift remediation and compliance packs are gated to paid editions. Check the state backend decision first: leaving Pulumi Cloud for a self-managed backend gives up ESC, Deployments and Neo, and no export path to another IaC tool is documented.

    Open questions and items to re-score

    • Dated GA announcements for Neo, the MCP server and Insights 2.0.
    • The CLI support and versioning policy page returned 404; no support window could be read.
    • Whether the Hyper-V package (pulumi/pulumi-hyperv, no releases) is usable; Hyper-V counts as uncovered.
    • The Xen Orchestra provider's release date and the Nutanix package's maintenance beyond one individual.
    • FIPS builds and air-gapped self-hosting of Pulumi Cloud Enterprise were not found.
    • Pricing before 2026-05 is not on record; the tracker's Free tier Neo allowance conflicts with the pricing page.
    • The 3.x release history between 2021 and 2026-08 was not retrieved; the yearly capability line between those dates is held flat.

    Related reading

    More from Infrastructure Automation Ledger 2026