Energy / ESG — carbon accounting software · Deep dive

Watershed

Enterprise carbon accounting software for the Fortune 500 — ~$108M ARR per Latka 2025, $1B+ valuation on a $100M Series C from Greenoaks in Feb 2024, now rebuilding its demand thesis as the Trump SEC rescinds climate disclosure and the EU Omnibus pushes CSRD into 2028.

emerging

The question that decides it: Does Watershed's four-year lead in ingesting Fortune 500 billing data, its ~$108M-ARR commercial momentum, and its multi-regulation compliance surface (SB-253, CBAM, surviving CSRD, voluntary SBTi) compound into a durable system-of-record for corporate carbon before the Trump SEC rescission of climate disclosure, the EU Omnibus delay of CSRD reporting to 2028, and SAP Green Ledger / Microsoft Sustainability Manager bundling natively into the ERP every one of Watershed's customers already pays for all three hit at the same time?

My take ★★ Very interesting ★ Interesting − Maybe ✗ Not interesting

Snapshot

Watershed is the carbon-accounting software company a wave of Fortune 500 CFOs picked between 2021 and 2024 when they decided they needed one — Stripe, Shopify, Airbnb, Walmart, FedEx, General Mills, Block, Spotify and Carlyle all appear on the public customer list. Founded in 2019 by three alumni of Stripe’s early climate team — Taylor Francis, Christian Anderson and Avi Itskovich — the company passed through Y Combinator’s Winter 2020 batch, raised a $70M Sequoia-led Series B at a $1B valuation in February 2022, and then a $100M Greenoaks-led Series C at a reported ~$1.8B in February 2024 per Bloomberg. Latka pegs 2025 ARR at ~$108M; Contrary Research and multiples.vc triangulate the business as the clear category leader in enterprise emissions software by both revenue and brand. The company’s defining problem for the next 24 months is not product: it is that two of the three regulatory tailwinds that justified the Series C mark — the US SEC climate-disclosure rule and EU CSRD on its original timeline — have been dismantled or deferred since the valuation was struck.

Founding story

The three founders met inside Stripe. Taylor Francis is the public face: a Yale undergraduate who co-founded Carbon Lighthouse in 2010 — one of the earliest energy-efficiency-as-a-service companies in the US, backed by Jamie Dimon’s family office among others — and spent the back half of the 2010s running operations there. In 2018 he joined Stripe, where he worked with Patrick and John Collison on what became Stripe Climate: the ability for any Stripe merchant to divert a fraction of each transaction toward carbon-removal purchases, launched October 2020. Stripe Climate seeded Frontier, the $1B advance-market-commitment vehicle Stripe, Alphabet, Shopify, Meta and McKinsey pledged in April 2022 to pull permanent carbon removal down a cost curve.

Christian Anderson joined Stripe out of a growth background and ran a series of launches inside the company before leaving to co-found Watershed. Avi Itskovich was a senior Stripe engineer with roots in Stripe Terminal and the payments stack. All three left Stripe in 2019 to attack the question Stripe Climate had surfaced inside every merchant conversation: what is my carbon footprint, measured against what accounting standard, with what data, and who audits it?

The initial product — the Y Combinator W20 pitch — was not what Watershed ships today. According to the Contrary Research company breakdown and multiple founder podcasts, the team spent roughly a year ingesting enterprise emissions data by hand for a dozen design-partner customers (Stripe itself the first) before productizing the ingestion-and-calculation engine. The pivot was less a change of market than a decision to productize a consulting workflow — a GHG-Protocol-aligned emissions inventory, built from a customer’s own billing and ERP data, delivered as a durable software system rather than a one-off deliverable. Stripe, Sweetgreen, Shopify and Patreon were the earliest customers. The Collison brothers both wrote cheques personally.

How it works

A Watershed deployment starts with data pipes. The customer sends spend-based and activity-based data from the systems that already run its business: utility bills, electricity meter data, cloud bills (AWS/Azure/GCP), T&E spend from Expensify or Concur, procurement spend from SAP or Oracle or NetSuite, freight-forwarder manifests from Flexport, and vendor-level invoice data by GL account. Watershed’s platform maps each of those line items against its emissions-factor library — the asset the company extended materially when it acquired VitalMetrics in May 2023, folding the CEDA (Comprehensive Environmental Data Archive) multi-regional economic input-output database into its calculation engine under the guidance of Dr. Sangwon Suh, VitalMetrics’ founder.

The output is a GHG Protocol-aligned inventory: Scope 1 (direct fuel combustion), Scope 2 (purchased electricity, with both location-based and market-based methods calculated), and Scope 3 — the fifteen upstream and downstream categories where most Fortune 500 emissions actually live. Scope 3 Category 1 (purchased goods and services) is where Watershed does the heaviest lifting and faces the heaviest criticism: a spend-based estimate ($/vendor $/category -> kgCO2e) is defensible as a first pass but not as the final auditable number, so Watershed built a supplier-engagement module that pushes a lightweight data-collection survey to the customer’s largest vendors, replacing the economic estimate with the vendor’s own reported activity data over time. Measure records are versioned and auditable; the Reduce module runs the inventory through intervention scenarios (switch to renewable PPAs here, reformulate this SKU, change this transport mode); the Report module exports to CDP, CSRD, SB-253/SB-261, CBAM filings and the TCFD framework. Watershed Marketplace is the carbon-credit storefront, with a 2026 RFP open for ~1 megatonne of removal credits sourced from Watershed-vetted suppliers (ESG News, 2025).

Product and business overview

The platform splits into four named components. Measure is the ingestion, mapping and calculation engine — the core SaaS line, and the asset the VitalMetrics acquisition most directly deepened. Report is the compliance surface: configurable templates and audit-ready outputs for CSRD, SB-253/SB-261/SB-219, CBAM, UK SECR, the SEC rules as proposed before rescission, SBTi targets, CDP responses and the Workiva export (per Workiva’s own documentation). Reduce (sometimes labelled Act) is the intervention and target-setting module: the place a sustainability lead builds a decarbonization pathway and tracks progress against SBTi 1.5°C-aligned trajectories. Watershed Marketplace is the storefront for high-integrity carbon removal credits, with a 2026 RFP open for roughly one megatonne of durable removal — direct air capture, enhanced weathering, biochar, biomass carbon removal and storage (BiCRS).

Two newer products matter for the next-twelve-months thesis. Watershed launched an AI-Driven Product Footprints feature in 2025 (ESG News) that generates product-level life-cycle estimates from procurement and bill-of-materials data — a direct answer to Carbonfact-style vertical LCA. And in May 2025 Watershed released a Free Global Emissions Database (The Financial Analyst) — a strategic concession to pressure from Normative and Climatiq’s open emissions-factor APIs. The free database is a loss-leader for Measure.

Business model and pricing

Watershed is annual, per-engagement SaaS with pricing custom-quoted. The published and triangulated figures as of 2026:

• Entry-level ACV ~$100,000 for a Fortune 1000 initial engagement (Vendr marketplace listing, 2026).
• Mid-market deployments $200,000-$500,000 blended ACV (Contrary Research breakdown; TrustRadius 2025 pricing summary).
• Enterprise deployments $500,000-$1M+ for Fortune 100 multinationals with twenty-plus-site Scope 1 inventories and complex multi-jurisdiction reporting (Vendr; dcycle.io pricing guide 2026).
• One-time implementation / data onboarding fees on top of the ACV; most multi-year deals bundle a professional-services envelope that reads to the customer as a managed-service engagement.
• Marketplace take-rate on carbon removal credit purchases run through the platform — undisclosed, but the economic model is a facilitation fee closer to a brokerage than a SaaS take-rate.
• Gross margins are not disclosed; Contrary Research and multiples.vc frame them as in-line with enterprise SaaS once implementation services are stripped out.

Attribute every ARR number to a date. Latka reports $108.3M ARR for 2025 and a (dated) $1B valuation — the latter trailing the $1.8B Series C mark and probably a conservative read; the Breathe ESG analysis of the Series C implied an ARR in the $50-80M range at close in Feb 2024, consistent with a double-digit multiple on 2x-ish YoY growth. Watershed is heavy at the top of the ACV distribution, which is why the business runs with ~500 employees (Revelio Labs 2026) rather than the ~1,500+ a per-seat SaaS at the same revenue would carry.

Traction over time

Market analysis

The carbon-accounting-software TAM is a function of regulation plus voluntary demand. Independent estimates diverge wildly — Fortune Business Insights’ 2025 report pegged the market at ~$18B by 2032; Grand View Research sized it in a similar band; the more promotional outliers (SNS Insider, Precedence) push to $100B+ by 2035 which should be read as directional only. The investable TAM for enterprise-grade platforms competing with Watershed is probably the $4-8B/year software-plus-services line by 2030, concentrated in roughly the 50,000 companies globally scoped into CSRD plus the 5,300 US entities scoped into California SB-253 (>$1B revenue doing business in California, per CARB’s own 2025 guidance).

Three structural forces define the next two years. (1) US retreat. The SEC voted in March 2025 to stop defending the Biden climate-disclosure rule in the Eighth Circuit, and in June 2026 the SEC proposed formal rescission (SEC press release 2026-49, Hunton, Duane Morris). The regulatory bid in the US goes from mandatory-disclosure demand to voluntary-SBTi demand — a smaller, slower buying motion. (2) EU ‘simplification.’ The Omnibus I package, adopted by the European Parliament in April 2025, deferred CSRD scope-expansion to wave-two companies by two years and compressed what counts as mandatory ESRS disclosure (Sidley, Skadden, Linklaters). The March 2026 Omnibus substance amendments (Morgan Lewis) further narrowed scope. The CSRD-driven TAM Watershed priced in at Series C is now roughly 60-70% of what Feb-2024 modelling assumed. (3) California and CBAM hold. SB-253 and SB-261 held through 2025 enforcement litigation — Watershed’s own California solution page quotes CARB’s November 2025 guidance; the EU CBAM mechanism, which taxes imported carbon-intensive goods, took its first scheduled compliance step in October 2026. These two are the mandatory-disclosure tailwinds Watershed is now most commercially levered to.

Competitive intel

The named rivals sit in the frontmatter. The important distinctions:

Persefoni is the one that mattered. The direct US peer raised ~$180M across its earlier rounds and traded as the premium alternative on PCAF-aligned financial-services use cases, but has visibly contracted — a $23M Series C extension in April 2025 (Fintech Global) at an undisclosed valuation, after a public pivot toward certified Chief Climate Officer training and asset-manager clients. Where Watershed broadened across Fortune 500 industrials and consumer brands, Persefoni concentrated into the one buyer segment — US climate-finance — that the Trump SEC retreat is most likely to shrink.

SAP Green Ledger is the platform risk. SAP was named a Leader in the IDC MarketScape 2026 for Carbon Accounting (SAP press release, May 2026). The Green Ledger thesis — carbon lives in the general ledger alongside financial GL, booked as actuals from invoiced activity rather than averages — is a direct architectural argument against Watershed’s spend-based estimates. Every Watershed customer running S/4HANA is a Green Ledger prospect, and SAP sells into procurement by bundling into existing RISE and GROW agreements.

Microsoft Sustainability Manager is sold on Azure consumption and integrates directly with Microsoft Fabric / Dataverse — the sustainability data model reads from Dynamics 365 and the Microsoft Cloud without the data-ingestion lift Watershed charges to perform.

Sweep is the European mirror image — named a Leader in the same Verdantix 2026 Green Quadrant as Watershed. If Sweep converts its CSRD-first positioning into US expansion, it is the one credible venture-scale peer attacking from Europe.

Normative, Plan A, Greenly, Carbonfact, Terrascope fill vertical and SMB edges Watershed concedes. Normative’s free SME calculator with the UN SME Climate Hub drives bottom-up adoption Watershed cannot afford to chase; Carbonfact owns apparel-vertical LCA; Terrascope has the Asia distribution Watershed does not; Greenly has a $60M-raised European SMB motion; Plan A has TÜV Rheinland CSRD pre-approval.

History and evolution

2019: Francis, Anderson and Itskovich leave Stripe to start Watershed. W20: Y Combinator batch. Oct 2020: public launch alongside Stripe Climate’s own public rollout; Shopify, Stripe and Sweetgreen are reference customers. Apr 2021: Series A led by Kleiner Perkins; the Collison brothers and Elad Gil ride along. Feb 2022: $70M Sequoia-led Series B at $1B — the moment Watershed clears the field commercially among US scaled startups. 2022-23: Fortune 500 land-grab (FedEx, General Mills, Walmart, Carlyle appear on the Trellis / customer-page roster). May 2023: VitalMetrics acquisition — the single most strategic move the company has made, bringing Dr. Sangwon Suh and the CEDA database inside. Jun 2023: Lowercarbon Capital investment. Feb 2024: $100M Series C led by Greenoaks at ~$1.8B — the peak-regulatory-tailwind mark. 2024-25: AI product footprints, free emissions database, big California SB-253 push. Mar 2025: EU Omnibus ‘stop the clock’ deflates the European CSRD demand curve mid-sales cycle. Jun 2026: SEC proposes rescission of climate-disclosure rule — the US side of the thesis also recedes. 2026: Watershed names itself Leader in Verdantix GQ 2026 and defends category position; the next-round question becomes whether the $108M ARR Latka reports in 2025 can grow enough through 2027 to defend a flat-to-up round at the $1.8B mark.

The visible stumbles: no US SEC rule to anchor the mandatory-disclosure story; the Omnibus delay through 2028; Persefoni’s down-round warning shot; the recurring G2/Gaia-comparison complaint that spend-based Scope 3 numbers are directionally wrong a third of the time (European Business Magazine’s ‘Carbon Accounting Is Wrong a Third of the Time’ is the most-cited external critique).

What people say

The case for. The praise clusters on three things. First, enterprise reliability — the Trellis profile of the $1.8B milestone lists FedEx, General Mills, Walmart, Carlyle, Shopify, Stripe, Airbnb, Block and Spotify as named customers, a Fortune 500 brand sheet no climate-tech peer matches. Second, methodology — the VitalMetrics acquisition brought Dr. Sangwon Suh (one of the world’s most-cited input-output LCA researchers) and the CEDA database inside; Verdantix and Forrester have named Watershed a Leader four separate times in the past two years per the company’s own analyst-recognition page. Third, product velocity — AI product footprints, the free emissions database, SB-253 solution page, Watershed Marketplace all shipped inside twelve months.

The complaints. First, Scope 3 estimation quality — the European Business Magazine ‘Carbon Accounting Is Wrong a Third of the Time’ critique explicitly names Watershed, Persefoni and Normative as representative of a category whose spend-based estimates diverge from LCA ground truth by double-digit percentages (Gaia Company’s comparison blog echoes the point). Second, pricing — Vendr, dcycle.io, Gaia and multiple RFP-wiki pages flag Watershed’s $100K-$500K ACV as prohibitive for sub-Fortune-1000 buyers and specifically contrast it against Greenly, Normative and Plan A at lower price points. Third, implementation burden — Hydrus.ai and Seedling.earth both note the long onboarding tail and heavy services envelope. Fourth, regulatory whiplash — the Watershed blog’s own ‘2025 Sustainability Whiplash’ post, intended as a reassurance, is the clearest acknowledgement that the US and EU regulatory stories have both softened. Glassdoor is 77-review-level signal, trending in the mid-4s overall with the usual pattern of start-up intensity complaints; no public mass-layoff event has been reported through Q3 2026, but climate-tech funding retrenchment since 2024 (CTVC, PitchBook sector reports) is the backdrop against which the next round will be priced.

Outlook: the open question

The answer conditions are these. For Watershed to be right over the next 24-36 months, four things have to be true. (1) The non-US mandatory stack — surviving CSRD scope after the Omnibus, California SB-253/SB-261, EU CBAM, UK SECR, SBTi 1.5°C targets — has to be large enough on its own to carry a $108M-ARR base to the $200M+ ARR a flat-to-up Series D or an IPO window would require. (2) The AI product footprints and free-emissions-database moves have to translate into bottom-up adoption that extends Watershed below its traditional $100K ACV floor without cannibalizing the Fortune 500 motion that pays the bills — i.e., Watershed has to execute a Figma-style self-serve wedge on top of its enterprise base. (3) The SAP Green Ledger and Microsoft Sustainability Manager bundles have to continue underperforming on data quality and emissions-factor coverage long enough for Watershed’s methodology moat (CEDA, Suh’s team, verified calculation engine) to compound into a durable system-of-record argument against them. (4) The Scope 3 estimation quality complaint has to be answered with a verifiable combination of supplier-engagement data and AI product footprints — not just better marketing, but defensible lower-uncertainty numbers audit firms will sign off on at reasonable-assurance level.

For Watershed to be wrong, the path is just as concrete. The Trump SEC rescission reduces US mandatory demand to zero and the voluntary SBTi/CDP tail is smaller than Series C modelling assumed. The EU Omnibus deferral pushes CSRD-wave-three to 2028 and compresses what gets filed in 2026-27 to a materially smaller surface. SAP Green Ledger closes the methodology gap at the Fortune 500 and bundles for free inside S/4HANA renewals. The next priced round comes at a flat-to-down mark, climate-tech private valuations re-rate, and Watershed’s $1.8B is the high-water line.

The single question that decides it: does Watershed’s four-year lead in ingesting Fortune 500 billing data, its ~$108M-ARR commercial momentum per Latka, and its multi-regulation compliance surface (SB-253, CBAM, surviving CSRD, voluntary SBTi) compound into a durable system-of-record for corporate carbon before the Trump SEC rescission of climate disclosure, the EU Omnibus delay of CSRD reporting to 2028, and SAP Green Ledger / Microsoft Sustainability Manager bundling natively into the ERP every one of Watershed’s customers already pays for all three hit at the same time? If yes, Watershed is the Workday of carbon and the IPO track opens on a $4-6B mark. If no, it is a very well-executed $100-200M ARR software company that got outflanked at the system-of-record layer and ends as a $1.5-2.5B strategic sale to Microsoft, SAP, Workiva or Workday.

How to attack it

The specific wedge is vertical-first Scope 3 calculation with physically-verified supplier data — not another horizontal carbon accounting platform. Carbonfact has already proven the shape of the attack in apparel: an LCA engine deeply embedded in a single vertical’s bill-of-materials data, where emissions are calculated from physical attributes (fibre type, dye process, factory electricity mix) rather than spend-based economic estimates. The same wedge is open-field in food & beverage (where General Mills, Nestlé, Kraft and the Big Four CPGs are all under SBTi FLAG and need activity-level Scope 3 that Watershed’s spend-based baseline cannot defend), in construction / cement / steel (where CBAM is already biting), in pharma (where contract-manufacturing emissions are opaque to the brand), and in logistics (where Flexport-style freight data is granular enough to replace estimates with actuals). A well-funded attacker ($30-50M seed-to-Series-A) would pick one vertical, build the physical-data pipeline — EDI feeds from suppliers, IoT meter data from factories, GLEC-framework-aligned freight manifests — and sell under Watershed’s floor by refusing to serve anyone outside the vertical.

The exploitable weaknesses in Watershed’s current position: (1) Scope 3 estimation quality is the recurring published complaint — European Business Magazine, Gaia, Hydrus and Seedling all cite spend-based inaccuracy specifically; a vertical attacker with physical data wins the methodology-defensibility argument. (2) Pricing floor at ~$100K ACV concedes the entire sub-Fortune-1000 denominator — Greenly, Normative, Plan A and Carbonfact collect it; a disciplined $20-50K/year SMB SKU with free emissions-factor APIs would hold the bottom of the funnel and the data that comes with it. (3) The regulatory bid is weakening on two of three fronts — SEC rescission (Semafor, June 2026) and EU Omnibus (Skadden, April 2025) both reduce the mandatory-disclosure demand that priced the $1.8B; an attacker whose GTM pitch is ‘you still need this even if nobody makes you’ (operational-cost savings, procurement risk, carbon-tariff exposure) carries a more durable buying motivation than compliance. (4) SAP Green Ledger and Microsoft Sustainability Manager bundles are both now shipping; a system-of-record argument is the one argument Watershed cannot win against a vendor whose software is already the system of record. (5) Services envelope is a cost-structure liability — Watershed’s ~500-person headcount against ~$108M ARR implies implementation-services drag; an AI-native competitor that automates ingestion with LLM-driven document extraction and supplier-portal scraping would operate at materially higher per-head revenue. (6) The methodology lead expires — CEDA is a 2024-2026 edge; open emissions-factor databases (Climatiq, Watershed’s own 2025 free database, OpenLCA, EcoInvent) are commoditizing the input layer, which collapses the methodology moat into execution and distribution over a 36-month horizon.

Adjacent-segment play

The most interesting adjacent-segment play is financed-emissions underwriting and procurement risk scoring — repackaging the same ingestion-and-calculation engine for a different buyer: not corporate sustainability leads, but the credit, procurement, and underwriting teams that already price supply-chain and counterparty risk. The data asset Watershed has built — four years of Fortune 500 Scope 1/2/3 inventories with vendor-level visibility — is the raw material a Moody’s-ESG-style scoring product needs, and the buyer is any CFO office, bank credit committee, or category-manager procurement org pricing supplier risk. Persefoni’s pivot toward asset-manager customers hints at this; a Watershed-grade financed-emissions underwriting product sold into regional banks, PE firms doing diligence, insurance underwriters pricing transition risk and procurement teams running vendor-scorecards would command a very different buying motivation (financial materiality, not CSR narrative) and would be insulated from the regulatory whiplash hitting the core SaaS line.

Second adjacent: public-sector audit tooling. EU CBAM authorities, state regulators (CARB in California), the UK FCA and the Monetary Authority of Singapore all need technology to verify filed disclosures. Watershed’s Report module is already an audit-ready export; a public-sector SKU priced into regulator budgets would be a non-obvious buyer segment with sticky multi-year contracts and no competition from SAP/Microsoft bundles.

Third adjacent, down-market: a vertical-SMB SKU at <$20K ACV, built on the free emissions database released in May 2025 and the AI product footprints engine. Does the physics of the wedge generalise? For the ingestion-and-calculation engine, yes — the pipeline reads the same billing data from a 50-person e-commerce brand as from a 50,000-person Fortune 500. The reason Watershed has not shipped this is channel conflict with its $100K+ enterprise motion, which is exactly the kind of blind spot an attacker is paid to exploit.

Sources and further reading

• Report: Watershed’s Business Breakdown & Founding Story — Contrary Research, 2024
• This $1.8 billion startup tracks carbon for FedEx, General Mills and Walmart — Trellis, 2024
• Watershed Climate Data Service Secures $100 Million Funding Round — Bloomberg, February 2024
• Watershed has raised a $100M Series C, at a $1.8B valuation — Watershed, February 2024
• Watershed lands $70M at $1B valuation — PitchBook, February 2022
• Watershed Revenue 2025: $108.3M ARR — Latka, 2025
• Trump administration rolls back climate disclosure rules — Semafor, June 2026
• EU Parliament Votes To Delay Implementation of Sustainability Reporting — Skadden, April 2025
• SAP Named a Leader in the IDC MarketScape: Worldwide Carbon Accounting 2026 — SAP, May 2026
• Climate Solutions Platform Watershed Acquires Carbon Data Provider VitalMetrics — ESG Today, May 2023
• Persefoni bags $23m funding to enhance carbon accounting software — FinTech Global, April 2025
• Unpacking Watershed’s $1.8 Billion Valuation: What It Means for Climate Tech — Breathe ESG, 2024