For Lena Thiede. About how the in-house LCA methodology you've run since Fund I becomes machine-inspectable per-portco config — defensible the day SFDR 2.0 enters force.
The European Commission's 20 November 2025 proposal restructures Article 9 into a regime demanding product-level PAI evidence. Your LCAs already are that evidence — but only if reconstructable at the portco line.
cases/planeta · audit_chain_v1 · methodology_overlay_v1
Two parallel observations. One artifact. One voice. One inflection that binds them.
Since Fund I close (Feb 2023), Planet A has run consequential LCAs in-house for every investment — ~50 pages per portco, published openly at planet-a.com/lcas. Two operational layers sit on top: the Improvement Rate vs reference, and the science-team veto. 50% of carry is tied to portfolio impact targets across the 4-area framework — climate, waste, resources, biodiversity.
Co-Founder and GP, leading the science team. First European VC to publish a biodiversity-impact methodology.
"We go beyond climate mitigation and also look at biodiversity protection, resource savings, and waste reduction."
- SFDR review proposal COM(2025) 841 final (20 Nov 2025) retires the Article 9 "Dark Green" label.
- Three new categories take its place — each with a 70% product-level PAI threshold.
- The next 12–18 months turn the published LCAs into the regulatory primary source.
Five dimensions sit under the LCA. Each is already a Planet A-published rule. The question SFDR 2.0 asks is whether each output is reconstructable from primary sources — per portco, on a 24-month look-back, by someone outside the science-team meeting.
FIG. ── PER-PORTCO EVIDENCE REGIME · 5 LCA DIMENSIONS × TODAY'S WORKFLOW vs UNDER SFDR 2.0
What changes is the reconstruction burden. Same calculations, three readers — regulator, LP, EU supervisor — all needing one source of truth.
Two primitives sit underneath the LCA methodology — audit chain and methodology overlay. The published PDFs stay public. The internal layer makes them survive 27+ portcos.
A hash-chained, append-only log of every per-portco LCA decision — reference product selection, lifecycle-stage emissions, Improvement Rate calculation, area weight assignment, science-team veto event. Once written, no entry is editable; corrections happen by additional entry, time-stamped, with the prior LCA version still readable.
Each entry carries an evidence-mode tag: [founder-reported] [planet-a-science-validated] [third-party-audited]. The science-team veto becomes a hash-chained event: which scientist signed off, when, against which reference product, with what data sources. The 50% carry-to-impact link runs against the chain, not against the published PDF.
Each portco's LCA inputs, reference product, lifecycle-stage emissions, Improvement Rate, area weights, and PAI exposures live as inspectable YAML — one file per portco, version-controlled, diff-able quarter over quarter.
As reference baselines shift, the Improvement Rate updates and the chain records the delta. The published LCAs remain the public external doctrine. The overlay is the internal layer that makes the doctrine survive 27+ companies and Fund II diligence.
Together, the AUDIT CHAIN names what evidence exists; the OVERLAY names what the evidence is supposed to show. Same source of truth, two different reading frames — one for the Commission auditor, one for the LP.
FIG. ── AUDIT CHAIN · LCA-DECISION BLOCKS · HASH-LINKED · APPEND-ONLY · NEVER-DELETABLE
FIG. ── 4-AREA FRAMEWORK · DETERMINISTIC ROLL-UP · NEW ARTICLE 9 SUSTAINABLE · 70% PRODUCT-LEVEL PAI
27+ portcos. Twelve to eighteen months until the LCA is the regulator's primary source. Pressure-test the regime.
The substrate answers in its own voice — the same primitive we'd install on one impact area, three portcos, one quarterly cycle. Not a demo. The actual reasoning surface.
STREAMING · κ ≥ 0.85 · ANONYMITY HOLDS · CITATIONS BY DEFAULT
One impact area. Three portcos. One quarterly cycle. One LP-reporting page per portco, generated from the chain.
If relevant, I can send a short async questionnaire — no call needed — and build a no-charge synthetic-data Phase 0 PoC "Planet A AI OS". Within two weeks, replay one portco's LCA evidence chain against the SFDR 2.0 product-level PAI regime before regulators start reading the published methodology as the primary source.
Article 9 carbon-math at €204M · LCA chains hardened as auditable substrate.
AI OS for the Urban Stack at €500M · climate adaptation across the built environment.
Published impact-carry formula meets non-linear evidence-chain growth — and an SFDR 2.0 replay.