SysML v2 Adoption Report: Early User Experience from 340 Practitioners

A survey of 340 systems engineers with SysML v2 pilot experience reveals: significant improvements in semantic consistency, mixed results on tool maturity, and an underestimated migration effort from v1 models and workflows.

SysML v2 in Practice: The Survey Data

Object Management Group ratified SysML v2 in late 2023 after a decade of development. Eighteen months into commercial availability, enough practitioner experience has accumulated to draw useful conclusions from structured survey data. A consortium of INCOSE chapters surveyed 340 engineers who had completed at least 3 months of SysML v2 pilot work on real projects.

Semantic Improvements Are Real

The survey's most consistent finding: practitioners who have used both versions rate SysML v2's semantic consistency significantly higher. The specific improvements most cited: the elimination of dual stereotypes for the same concept (a persistent v1 frustration), the cleaner separation between structural and behavioural models, and the KerML kernel's explicit semantic foundation.

73% of respondents agreed that v2 models were easier to check for consistency than equivalent v1 models. This aligns with the language design goals: KerML's formal semantics enable automated consistency checking that was impossible in v1's informally-specified metamodel.

Tool Maturity Is the Constraint

The consistent blocker: tooling. Unlike v1, which had mature commercial tools from Sparx, PTC, IBM, and No Magic by the time most practitioners adopted it, v2 launched into a tooling ecosystem that was (and remains) catching up with the specification. At time of survey, respondents identified three primary tools in active use; all had capability gaps relative to v1 equivalents; and interoperability between tools was rated 2.1/5.

Respondents on vendor roadmaps estimated tool parity with v1 capabilities by mid-2026 — consistent with current commercial tool development trajectories.

Migration Effort Is Underestimated

The most consistent surprise: migration effort from v1 to v2 was significantly higher than teams anticipated. The language is not backwards-compatible; v1 profiles and stereotypes do not map cleanly to v2 library packages; and v1 tooling export formats are not importable into v2 tools. Organisations with large v1 model repositories face a genuine migration project, not a tool upgrade.

Recommendation from high-satisfaction respondents: treat SysML v2 as a new start for new programmes, and maintain v1 for existing programmes until the migration cost-benefit is clearly positive.