SINE Strategy Wiki

PACT / Pathfinder Network

20 min read

The Partnership for Carbon Transparency (PACT) is a global initiative launched in 2020 and hosted by the WBCSD, to build an aligned, open approach for calculating and exchanging full product carbon footprints (PCFs). It started as a coalition of 60+ leading companies launched at COP26 (2021) — the flagship use case in the 2022 Data Commons whitepaper — and has since grown to 2,500+ companies engaged globally, with participation from 150+ stakeholders across 11+ industry-specific initiatives, per PACT Methodology v3.0 (WBCSD, 2025).

⟨Note: this page originally documented PACT only through the secondary lens of the 2022 Data Commons whitepaper. Per Martin Pompéry, the wiki’s documentation task extends beyond Data Commons — this page (and its companions below) now also stand as direct technical/organizational reference on PACT, independent of its role as a Data Commons use case.⟩

The methodology and data-quality/verification mechanics are documented in depth on dedicated pages — this page covers organizational identity, the network layer, and the Data Commons framing:

Why it exists

Scope 3 emissions (all indirect emissions in a company’s value chain, upstream and downstream) are the main emissions driver for most companies (“your emissions are my liability”), and full transparency across complex global value chains can’t be solved by any single company acting alone.

What PACT built

  1. Pathfinder Framework — a methodology that harmonizes and standardizes how PCFs are calculated, removing room for interpretation/inconsistency in existing methods, plus rules for data exchange and a quality-assurance/verification scheme.
  2. Pathfinder Network — built with SINE Foundation, on open-source technology and open standards, explicitly to remove vendor lock-in and protect data sovereignty (decentralized approach + strong encryption) while staying open/interoperable across the different technology solutions companies already use.

⟨2026-10-29: PACT’s own founding paper names three “Pathfinder” focus areas, not two — Framework, Network and Pathfinder Ecosystem, the last being its alignment work with industry initiatives, disclosure organizations, standard-setting bodies and regulators. See the Vision Paper. The Ecosystem pillar has no page here; the source defines it only by activity.⟩

PACT introduced the Primary Data Share (PDS) — a KPI businesses must disclose for every data exchange (mandatory now, not deferred — see PACT Data Quality & Verification for the exact “shall” language and formula), tracked network-wide with a stated goal of pushing it toward 100%, i.e. rewarding primary-data use over estimates.

The PACT Network (technical layer)

Software-technology layer for exchanging standardized PCF data: an open, global, peer-to-peer network of interoperable, “PACT Conformant” solutions for secure exchange of accurate, primary, and verified PCF data across industries and value chains. Since the public release of the PACT Technical Specifications v2 (February 2023), 40+ solution providers across 14 countries have implemented the specifications and become PACT Conformant — able to both calculate PCFs per the PACT Methodology and exchange them over the network. (PACT Conformance certifies interoperable exchange; it is a separate claim from following the PACT Methodology for calculation itself, though many providers do both.) The companion PACT Technical Specifications v3 / Simplified Data Model 3.0 define the exact data attributes exchanged (external, at docs.carbon-transparency.org — not itself ingested).

⟨Lint 2026-10-05: read the adoption figure above together with the v2 sunset below. It counts conformance against a specification family that has since been retired, the sentence was written before the sunset was known, and it still reads present-tense. How many of the 40+ have moved to v3 is not recorded anywhere in the wiki. Note also that two different artefacts in this wiki are called “v2” — the PACT Technical Specifications v2 (February 2023, sunset 01.04.2026) and the PACT Methodology v2 (2023, the former “Pathfinder Framework”, superseded by Methodology v3). The sunset date attaches to the first. Nothing on record says whether the two retirements are the same event.⟩

What the protocol actually is ⟨2026-11-06⟩

⟨Added from the Technical Specifications for PCF Data Exchange v2.2.0 (released 10 April 2024) — the first time the wire format itself is in the wiki rather than described second-hand.⟩

The Tech Specs are an interoperability profile ⟨Per Martin Pompéry, 2026-11-06⟩. The protocol invents almost nothing: it selects from existing standards and constrains them — OAuth2 Client Credentials and Bearer tokens, OpenID Connect discovery, a named subset of the OData v4 $filter grammar, CloudEvents in Structured Content Mode, RFC 8288 Link pagination, URN namespaces over GTIN/CAS/InChI, UN CPC, ISO 3166, SI units, UUIDv4. Its own §7.1: “The scope of the HTTP API is minimal by design.”

Its governing design principle is Postel’s law ⟨Per Martin Pompéry, 2026-11-06; the term does not occur in the document, the behaviour is written throughout⟩. A mandatory floor — HTTPS only, four actions (Authenticate, ListFootprints, GetFootprint, Events), server-side pagination — with liberal acceptance above it: a host “should strive to accept an event”; filtering is optional and may be ignored in whole or in part; HTTP 202 returns deliberately incomplete results; senders must retry with exponential backoff and “should not rely on” the receiver responding.

⟨Inference: that is what lets forty-odd independently built host systems interoperate without upgrading in lockstep, and it is the correct frame for reading everything the specification does not contain. A profile’s substance is which existing pieces it picks. Which makes the one place it declined to pick — see the security inventory below — a decision rather than an omission.⟩

§7.2.1 names what is out of scope, and the list is the answer to the Vision Paper’s open questions: footprint calculation; authentication and access management (“the act of deciding and setting which product footprint may be accessed by each data recipient”); credentials management; logging (“access logs and audit trails”). Conforming host systems are called solutions, and solutions “add further functionality on top of this standard”. ⟨Inference: two conformant hosts can differ completely in who may see what, on what evidence, with what audit trail. The ICC’s “exchanges data rather than accountable quantities” is, read against §7.2.1, a description of the scope statement.⟩

SINE is inside the specification, not just beside it. Comments go to public-dev@pathfinder.sine.dev; every normative API example in §7.10 runs against host api.pathfinder.sine.dev and returns server: Pathfinder.

Why it’s held up as the model case

PACT’s own values map directly onto the Data Commons three values: equality/inclusivity ≈ Reciprocity, data confidentiality/sovereignty ≈ Security, openness ≈ Openness. ⟨2026-10-29: the primary source has five design principles, not three — Data confidentiality and sovereignty, Data quality and integrity, Scalability and diversity, Equality, Openness (Vision Paper). The three-value mapping is a compression performed by the Data Commons whitepaper, not by PACT. Equality ↦ Reciprocity and Openness ↦ Openness are clean; Security absorbs two principles; Scalability and diversity has no counterpart in the triad.⟩ ⟨2026-11-01: and the value set itself moves — see The moving value set below.⟩ The whitepaper credits this alignment — plus a collaborative development process where corporations and software providers built the network’s technical core together, rather than one party dictating it — for PACT achieving its first cross-solution PCF data exchange just months after launch, on a timeline the paper says would normally take years given the number of stakeholders involved.

The moving value set (2022 → 2023)

⟨Added 2026-11-01 from Pathfinder Framework v2, Figure 19 — recovered from a page image after Marker garbled the table.⟩

Seven months after the Vision Paper, PACT restated its values. The list is six “key values”, not five design principles, and the membership changed:

Vision Paper, June 2022 (5 design principles)Framework v2, Jan 2023 (6 key values)Data Commons, 2022 (3 values)
Data confidentiality and sovereigntyData confidentiality and sovereignty 🔒Security
Data quality and integrity— absent —(absorbed into Security)
Scalability and diversityScalability—
EqualityEqualityReciprocity
OpennessOpennessOpenness
—Inclusivity (new)(≈ Reciprocity)
—Reach (new)—

Two readings, both recorded:

  • ⟨Inference: Data quality and integrity did not vanish from PACT — it became the Framework’s subject matter. A guidance document about PCF calculation, PDS, DQR and assurance does not need to list data quality as a value; it operationalises it as requirements. The June 2022 list was aspirational; the January 2023 list sits in a document that delivers the quality half.⟩
  • ⟨Inference: the two additions, Inclusivity and Reach, are both about membership and alliance rather than about the data. They appear in the version whose collaborator circle grew from 38 to 50 and which explicitly courts “synergies with other initiatives”. The value set shifted from describing a system’s properties to describing a partnership’s conduct.⟩

The v2 definitions as printed:

ValueDefinition
InclusivityAll members jointly shape the Partnership by providing expertise, resources, data, and active participation in targeted working groups
EqualityThe Partnership reflects the different ambitions of its members and ensures no member on their own is able to direct the initiative down a particular path
Data confidentiality and sovereigntyData owners have the flexibility to exchange different levels of data granularity with adjustable access rights and can approve each exchange
OpennessMethodological and technological solutions are designed to allow for interconnectivity, interoperability, and integration
ScalabilityThe Partnership creates an interoperable network of networks compatible across sectors and supply chains, serving multiple business needs
ReachSynergies with other initiatives focused on emissions transparency are fostered and leveraged as much as possible

Note the drift in the confidentiality value itself. In June 2022 it carried the strong claim — data that “protects organizations from any possible back-tracing of product-specific information”. In January 2023 it is defined as “adjustable access rights” plus per-exchange approval: an authorization model. The padlock icon printed next to it is the only cryptographic signifier in the document. See below.

What was promised in 2022, and what was built

⟨Added 2026-10-29 from the Pathfinder Network Vision Paper (June 2022), the earliest primary PACT source in this wiki. It was written before the Network existed — “the Pathfinder Network is still under development” — which makes it usable as a baseline rather than as a description.⟩

Figure 4 of that paper maps the planned user journey onto six mechanisms and marks which were already specified. Exactly one was: the data exchange mechanism. Identity management, authentication, plausibility and error-checking, and security and fraud prevention were stated intent. Auditing was placed under “Add-on features”, outside the core journey, by PACT’s own layout.

The descriptions attached to the unbuilt mechanisms were not modest. Authentication was to provide “the ability to trace, track and audit [PCFs] and their underlying accounting methodologies”; plausibility checking was to enable “full trust in the data once part of the Network”; the auditing mechanism was to support “ex ante and ex post verification processes […] in order to establish full transparency on and trust in the application of carbon accounting methodology and correctness of data.”

Against that baseline, the record this wiki holds: the data exchange mechanism shipped (Technical Specifications v2, February 2023; 40+ conformant providers), and verification is still a two-horizon roadmap in Methodology v3 (2025) — see PACT Data Quality & Verification.

⟨Inference: the four charges in the ICC’s Table 4 below land on the boxes PACT itself left white in June 2022. That reframes them. “Exchanges data rather than accountable quantities” is not a drift away from PACT’s design and not only an interested critic’s framing — the trust layer was scoped from the start, scheduled, and has not arrived in three and a half years, while the one mechanism that was ready at vision time became the thing PACT is known for. What this source cannot tell us is why: difficulty, absent demand, or governance. Open.⟩

The architectural exclusion

The same paper draws a boundary that the later debate keeps crossing:

“In contrast to individual technology solutions or platforms — which, for example, can support organizations with their emissions data management, including storage, calculation, visualization and analysis — the Pathfinder Network cannot be used to calculate a carbon footprint or to store any individual footprint data.”

With three companions: peer-to-peer “with no need for a central intermediary”; “Is PACT setting technological standards? No.”; and PACT “may in the future operate minimal network infrastructure strictly necessary for […] our mission.”

⟨Inference: no storage, no calculation, no central intermediary, not a standards body is a deliberate architectural commitment made before the first line of the Network was written. The ledger architecture proposed by the ICC in 2026 needs conserved quantities that persist and reconcile — the stateful layer PACT ruled out of its own scope. So the ICC is not pointing at an unfinished PACT; it is proposing a different system in the space PACT declared out of bounds. For open task 1 that is the sharper question: SINE’s response is not “we will add that”, it is either a defence of the boundary or a decision to move it.⟩

Confidentiality, stated without cryptography

The 2022 paper states the requirement at full strength — the “data-sharing dilemma”, each organization as “sole owner […] with full data control over who has access to any data point”, and data that “protects organizations from any possible back-tracing of product-specific information”. Its instruments are a governance framework, legal frameworks, identity management and fraud prevention, and data minimisation (“only the minimum necessary amount of data is exchanged”). Cryptography, encryption, zero-knowledge, MPC and proof do not appear in the document.

This is a fourth instance of the pattern in open task 19, and the one hardest to explain by unfamiliarity: SINE is the named technology partner on the document, and SINE’s first project (March 2021, with Cozero) was a secure-MPC carbon benchmarking tool.

The same finding, sharper, in Framework v2 (January 2023) ⟨added 2026-11-01⟩. A grep across all 78 pages for encrypt | cryptograph | zero-knowledge | privacy-enhancing | multi-party | anonymi | pseudonym | differential privacy | secure comput returns zero hits. “Confidential” appears three times, all in §6.2 (“straightforward and confidential cross–value chain and cross-industry exchange”); §1.3 adds “the secure exchange of product-level emissions data”. Neither is operationalised.

The one technical control in the document is Appendix B’s “Digital record signature” — “Digital signature covering the full PCF … Mandatory: No”. Optional, and it addresses authenticity, not confidentiality. It does not appear in Methodology v3 at all (grep for signature: no hits).

That makes this the fifth instance of open task 19, and the most precise: the gap is now documented across three consecutive PACT primary sources (2022, 2023, 2025), with SINE named as technology partner on each. SINE’s own self-description inside the v2 document is “ready-to-use governance tools and open-source software” — not PETs. ⟨Inference: this favours a sequencing reading of open task 19 — the confidentiality layer as a later ambition — over a “the technology wasn’t known” reading, since SINE both knew the technology and described itself as offering something else.⟩

Sixth instance, and the sharpest: the protocol itself ⟨added 2026-11-06 from the Tech Specs v2.2⟩. Everything the wire format provides, exhaustively: TLS, OAuth2 client credentials, Bearer tokens, optional OIDC discovery, and a check of the source field on an incoming event. A ProductFootprint travels as plain JSON; companyName is free text; the assurance object is a set of claims about an audit with nothing attached a recipient could verify. Confidentiality is access control, and even that is advisory — hosts “SHOULD implement an access management system”. Cryptography, encryption, signature, proof, zero-knowledge and multi-party computation do not occur in 2,421 lines. ⟨Inference: this instance cannot be read as the technology being out of view, because the document’s entire method is profiling other people’s standards. JWS was available on exactly the terms OData and CloudEvents were, and was not selected. Note too that the optional “Digital record signature” of Framework v2 (Appendix B, above) has no counterpart in the wire format at all — the field does not exist.⟩

And disclosure is discretionary by specification. §4.1.2: when a recipient requests a PCF it cannot see, “the data owner decides on making the PCF available or not”; when the PCF does not exist, “the decision making and decision making protocol for this case is up to the discretion of each data owner”. Refusal is a specified outcome (AccessDenied, HTTP 403) and leaves no aggregate trace.

Contradiction, held open. The 2022 Data Commons whitepaper describes the Pathfinder Network as protecting sovereignty via a “decentralized approach + strong encryption” (quoted at the top of this page). PACT’s own paper, two months later, describes the same protection as governance plus minimisation. Both are in sources/. Not resolved.

The external assessment (ICC, 2026)

The first assessment of PACT in this wiki that does not come from PACT, from WBCSD, or from a SINE whitepaper: The Carbon Accounting Landscape Review, published by Carbon Measures.

Read it with its premise in view — the Review was commissioned on the stated premise that a ledger-based architecture is the answer, and every “open item” it records is distance from that target, not a defect. It repeats this qualification four times. It is also, at the same time, an interested party’s account of why a competing architecture is needed.

What it credits PACT with. Status: “Scaling — rapidly expanding adoption among multinational supply chains; emerging as the leading interoperability and data-exchange framework, with active industry pilots” (§5.1). In §6.1, among the foundations the landscape already supplies: “PACT demonstrates that interoperable, supplier-specific product carbon data can be exchanged across multi-tier supply chains at scale, with verification — a practical answer to the systems-layer problem identified in Section 2.” PACT’s Primary Data Share is named, alongside CBAM’s tightening defaults, as one of the “two complementary mechanisms — disclosure and consequence” driving the shift to primary data. PACT is also one of the eight sources the Review’s own criteria are distilled from, credited with adding harmonisation and scalability to the ISO/GHGP principles (§4.1).

What it records as open. Table 4 (p. 32), verbatim:

“Exchanges data rather than accountable quantities; period-average static figures rather than transaction-level records; does not provide conservation, transfer, or reconciliation mechanisms; inherits the quantification latitude of its methodological basis.”

Appendix A, more bluntly: “Standardised cradle-to-gate exchange with primary-data and data-quality metrics; no verification emissions transfer as accountable quantities.”

The four charges, unpacked:

  1. Data, not accountable quantities. A PCF moves as an estimate attached to a message, not as a quantity that is conserved when it moves. Nothing in the exchange prevents the same emission being counted by two parties or dropped by both.
  2. Period averages, not transactions. PACT figures are static and period-scoped; a ledger records per transaction, aligned to financial periods.
  3. No conservation, transfer or reconciliation. PACT cannot reconcile a product figure to an entity total without gaps or overlaps — the Review’s criteria F (Conservation) and E (Fungibility).
  4. Inherited latitude. PACT’s comparability is bounded by the flexibility of ISO 14067 / GHGP beneath it. PACT narrows the boundary (fixed cradle-to-gate) but does not close the allocation optionality — see PACT PCF Calculation Methodology.

The division of labour it proposes (§5.3): quantification science from the ISO/GHGP family, sharpened by PEF’s prescriptiveness and ISCC CFC’s assurance discipline; exchange infrastructure from PACT; accounting architecture from E-Ledgers. “What none of these layers was designed to provide […] is the accounting architecture.”

⟨Inference: for SINE the assessment is a promotion and a demotion at once. PACT is named the leading exchange framework — that is external validation of the layer SINE co-built. It is also placed below the layer where the next standards fight will be fought, by a coalition that intends to define that layer. Whether the ledger layer stays interoperable with the exchange layer, as the Review promises, or eventually absorbs it, is the open question. This wiki does not resolve it.⟩

Contradiction held open, not resolved. The 2022 Data Commons whitepaper treats PACT as the validating success case for the Data Commons framework; the ICC Review treats the same facts as a well-executed partial layer. Both readings rest on the same evidence and differ on what was being attempted.

Open task: SINE has not yet responded to the architecture candidate. See open tasks.

The v2 sunset, and iLEAP’s departure ⟨2026-10-01⟩

From a source outside the PACT/WBCSD orbit — the iLEAP Technical Specifications v1.1.0-draft.2, §7.1: “Since April 1 2026, the v2 family of the PACT Technical Specifications, including [PACTDX], is sunset.” This is the first date in the wiki for the retirement of the v2 specification family — the generation whose February 2023 release produced the “40+ conformant solution providers” figure above.

⚠ What kind of evidence this is. ⟨Lint 2026-10-05: everything in this section comes from a document that is not a finished standard. v1.1.0-draft.2 is tagged a Living Document — iLEAP’s own lowest publication stage, below Consultation Draft and below Release — and the last stable iLEAP release is v1.0.0 (2025-05-27). The source summary and the iLEAP page both say so plainly; this section, written the same day from the same source, dropped the qualifier and states the decoupling as accomplished fact (“has stopped riding on PACT”). Two consequences, which differ in strength and should not be merged: the sunset date is a report about someone else’s published lifecycle and a draft can report that accurately, so it is probably reliable — but it is uncorroborated and PACT/WBCSD sources in this wiki do not mention it. The removal in v2.0.0 and mandatory standalone API are iLEAP’s own forward intentions in an unreleased draft, and those can still change before Release. Treat the first as likely, the second as announced-not-settled.⟩

The visible consequence: iLEAP has stopped riding on PACT. Until v1.1.0 iLEAP data travelled as a data-model extension inside a PACT ProductFootprint, over the PACT Data Exchange Protocol. v1.1.0 introduces a standalone iLEAP API, makes it mandatory for conformance, keeps the PACT path only “for backward compatibility and fallback purposes”, and schedules it for removal in v2.0.0. Semantic interoperability with PACT is now stated as an aspiration for future releases, not a property.

⟨Inference: the cause is PACT’s own release management rather than any disagreement on substance, and the mapping tables are retained. But the practical effect is that the two standards SINE co-built no longer share a wire. Whether the “Internet of Emissions Data” converges on one network or settles into sector-specific networks joined by mapping tables is an open strategic question that neither document addresses. Worth putting to Tell Münzing.⟩

Cross-references

Sources

  • PACT, June 2022) — sources/standards/2026-10-29-pact-network-vision-paper/2026-10-29-pact-network-vision-paper.md (Figures 1, 4, 5; design principles; FAQ)
  • sine-data-commons-2022 — sources/whitepapers/2022-04-25-data-commons-final.pdf
  • PACT Methodology v3.0 (WBCSD, 2025) — sources/whitepapers/2026-08-10-pact-methodology-v3.0/pathfinder-methodology-v3.0.md
  • ICC Carbon Accounting Landscape Review (2026) — sources/whitepapers/2026-08-17-icc-carbon-accounting-landscape-review/2026-08-17-icc-carbon-accounting-landscape-review.md (§4.1, §5.1, §5.3, §6.1, Table 4, Appendix A)
  • iLEAP Technical Specifications (2026) — sources/standards/2026-10-01-ileap-technical-specifications-v1.1.0-draft.2.md (§1.2, §7.1 — the v2 sunset and the decoupling)
  • PACT, Apr 2024) — sources/standards/2026-10-29-pact-data-exchange-protocol-v2.2.md (§1.1, §4.1, §5, §6, §7.1–7.10, Appendices A–B)
  • PACT, Jan 2023) — sources/standards/2026-10-29-pact-methodology-v2-framework-2023/…md (Figure 19 key values, recovered from page image; §1.3, §6.2; Appendix B; back matter)

Graph View

Backlinks