Atmospheric AI
Computes direct materials value chains facility by facility to inform sourcing decisions on cost, risk, and carbon.
Website: https://atmosphericai.com/
Cover Block
Open sources
| Field | Value |
|---|---|
| Company | Atmospheric AI |
| Tagline | Computes direct materials value chains facility by facility to inform sourcing decisions on cost, risk, and carbon [Atmospheric AI] |
| Headquarters | London, UK [Crunchbase] |
| Founded | 2025 [Crunchbase] |
| Stage | Seed [Axios Pro, November 2026] |
| Business model | SaaS |
| Industry | Logistics / Supply Chain |
| Technology | AI / Machine Learning |
| Geography | Western Europe |
| Growth profile | Venture Scale |
| Founding team | Co-Founders (2), Roee Goldberg and Toby Darbyshire [DevCuration, November 2026] |
| Funding label | Seed |
| Total disclosed funding | ~$10.1 million [Axios Pro, November 2026] |
Links
Open sources
- Website: https://atmosphericai.com/
What an Investor Needs First
PUBLIC Atmospheric AI builds software that models raw-material supply chains facility by facility, and it merits attention now because investors have started to fund a broader procurement thesis around computed cost, risk, and carbon intelligence rather than supplier-reported data alone [Axios Pro, November 2026] [Atmospheric AI] [The SaaS News, November 2026]. Founded in 2025 and based in London, the company is early, but the initial narrative is unusually specific: it began with petrochemicals and plastics, where sourcing decisions are exposed to volatile input costs, opaque capacity, and emissions scrutiny [Axios Pro, November 2026] [The SaaS News, November 2026].
The core product is described as a computational model of raw-material production and distribution, with company materials summarizing the approach as computing direct materials value chains facility by facility; reporting says this helps large manufacturers assess production costs, capacity, supply risk, and emissions without relying solely on supplier-provided information [Axios Pro, November 2026] [Atmospheric AI]. Public materials also point to a customer-facing platform called Cortex, though the product naming is less consistently corroborated than the underlying use case, so the stronger verified point is the supply-chain modeling layer itself [freehire.me] [Axios Pro, November 2026].
The team has at least one founder with prior scaling exposure. Roee Goldberg previously co-founded OpenWeb and served as COO there, according to Axios Pro, Markets Insider, and LinkedIn-sourced profiles, while Toby Darbyshire is identified in seed-round coverage as a co-founder [Axios Pro, November 2026] [Markets Insider, December 2021] [LinkedIn, retrieved 2026] [DevCuration, November 2026]. Atmospheric AI has also added commercial weight with founding chief commercial officer Torsten Lichtenau, reported to have spent 22 years at Bain & Company, most recently as senior partner and global head of carbon transition [nstartup, November 2026] [Startup Mag, November 2026].
On capitalization, the clearest public financing event is a $10.1 million seed round announced in November 2026 and led by DBL Partners, with participation from Raise, Broadscale, Working Capital Fund, Tiny.vc, Norrsken Evolve, and Anamcara; the business model is presented as SaaS [Axios Pro, November 2026]. An August 2026 entry for a $6.3 million seed appears in the source set, but the November reporting reads more like the definitive disclosed round, so investors should treat the earlier figure cautiously until the cap table and round chronology are reconciled [Funding Spotter, August 2026] [Axios Pro, November 2026].
Over the next 12 to 18 months, the key items to watch are whether the company converts its initial footprint into repeatable enterprise deployments, broadens beyond plastics into chemicals, paper, and metals, and substantiates early traction claims such as working with about 100 manufacturers and modeling more than 50,000 industrial facilities worldwide [The SaaS News, November 2026] [Axios Pro, November 2026] [LinkedIn, retrieved 2026]. Haleon's reported multiyear work with the company on resin sourcing is encouraging, but the central diligence question remains whether Atmospheric AI can turn a technically ambitious modeling system into durable procurement workflow adoption inside large manufacturers [Startup Mag, November 2026] [Axios Pro, November 2026].
Partially corroborated -- Core company, funding, and founder-background claims are corroborated by multiple public sources, but several traction and product-detail claims remain single-source or partially corroborated.
Taxonomy Snapshot
| Axis | Value |
|---|---|
| Stage | Seed |
| Business Model | SaaS |
| Industry / Vertical | Logistics / Supply Chain |
| Technology Type | AI / Machine Learning |
| Geography | Western Europe |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (2) |
| Funding | Seed, total disclosed ~$10,100,000 |
Inside the Company
PUBLIC
The public record on Atmospheric AI is still thin, which is typical for a newly launched seed-stage company. Its own website describes the company as computing direct materials value chains facility by facility to inform sourcing decisions on cost, risk, and carbon, while Crunchbase lists it as a London-based company founded in 2025 [Atmospheric AI, retrieved 2026] [Crunchbase].
That matters because the company appears to be framing itself less as a generic procurement tool and more as a system of record for how raw materials are actually made and moved. The same website positioning emphasizes "computed ground truth," a useful clue to the founding angle even though the company has not published a detailed origin story or legal entity information in the cited public materials [Atmospheric AI, retrieved 2026].
On milestones, the public company-controlled footprint remains limited. Crunchbase records Atmospheric AI as a founded company with funding activity, and the website provides a live product-level positioning statement, but neither source offers a fuller chronology of launch milestones, incorporation details, or operating history beyond the 2025 founding marker [Crunchbase] [Atmospheric AI, retrieved 2026].
Partially corroborated -- Confirmed by the company website and Crunchbase, with chronology details still sparse in those public sources.
Under the Hood
MIXED Atmospheric AI’s product claim is narrower, and more interesting, than a generic procurement AI pitch. Public reporting describes the software as a system for computing raw-material supply chains facility by facility, using computational models, effectively digital twins, of production and distribution rather than relying only on supplier-reported inputs [Axios Pro, November 2026] [Atmospheric AI] [The SaaS News, November 2026]. In plain terms, the company is trying to answer what a direct material should cost, where it is produced, how supply routes connect, and what cost, capacity, disruption, and carbon exposures sit inside that network [Axios Pro, November 2026] [Startup Mag, November 2026].
The initial product scope appears focused on petrochemicals and plastics, with expansion into chemicals, paper, and metals described in funding coverage rather than inferred here [Axios Pro, November 2026] [The SaaS News, November 2026]. A job-posting source and company-adjacent listings identify the customer-facing platform as Cortex, positioned as a B2B analytics product for materials cost, carbon, and risk, though the naming evidence is weaker than the broader product description and should be treated accordingly [freehire.me] [The SaaS News, November 2026].
The technical picture is partly visible from what the company and its hiring footprint have said in public. Atmospheric AI says it computes direct materials value chains facility by facility, while third-party coverage adds that the system models more than 50,000 industrial facilities worldwide and incorporates plant technologies, efficiencies, energy mixes, and material flows across shipping routes [Atmospheric AI] [Axios Pro, November 2026] [LinkedIn, retrieved 2026]. A surfaced engineering role points to Next.js, TypeScript, Python, and FastAPI as part of the application layer and data workflow, which is best read as stack evidence rather than proof of deeper model architecture (inferred from job postings) [freehire.me].
The main analytical appeal is that this is decision software built around computed industrial ground truth, not a lightweight reporting layer, but the public record still leaves important gaps. There is no verified public demo in the materials reviewed here, and several more expansive feature descriptions tied to scorecards, workflows, AI assistants, or broader engineering intelligence come from lower-authority directories and review-style sources rather than primary company materials or top-tier reporting [Gartner Peer Insights] [G2] [GeniusFirms, October 2026]. For public-market readers, the core product claim is credible enough to track, but the exact depth of automation, model refresh cadence, and integration burden remain open diligence questions.
Partially corroborated -- Core product positioning is corroborated by Atmospheric AI’s website and Axios Pro, while product naming and some feature detail rely on weaker third-party sources.
Market Research
PUBLIC
The market matters now because manufacturers are being asked to make sourcing decisions that satisfy three constraints at once, cost, resilience, and carbon, and those variables are increasingly moving together rather than separately [Axios Pro, November 2026] [Startup Mag, November 2026]. Atmospheric AI is positioned around direct materials procurement rather than general supply-chain software, which narrows the lens but also points at a budget line that tends to matter in periods of input volatility [Axios Pro, November 2026]. The available source set does not provide a third-party TAM, SAM, or SOM for Atmospheric AI’s exact category, so the public evidence supports a category read rather than a quantified market model.
What is visible from the reporting is a specific operating pain point. Atmospheric AI says it computes raw-material value chains facility by facility and models production costs, capacity, supply risk, and emissions, with an initial focus on petrochemicals and plastics [Atmospheric AI] [Axios Pro, November 2026]. That framing implies the company sits at the intersection of procurement analytics, industrial data, and sustainability decision support, a useful distinction because buyers in those domains often exist already inside large manufacturers even when a standalone “materials intelligence” budget does not [The SaaS News, November 2026].
The clearest demand driver is that supplier-provided information appears insufficient for the use case Atmospheric AI is targeting. Axios reported that the product is meant to help large manufacturers understand production economics and risk without relying solely on supplier disclosures [Axios Pro, November 2026]. Startup Mag also reported that Haleon had worked with the company for almost two years on resin sourcing by November 2026, which, if accurate, suggests that at least some strategic buyers are willing to test external computation as a complement to incumbent sourcing workflows [Startup Mag, November 2026].
A second tailwind is scope expansion across adjacent materials categories. The company’s starting point is plastics and petrochemicals, but multiple reports say it plans to extend into chemicals, paper, and metals [Axios Pro, November 2026] [The SaaS News, November 2026] [DevCuration, November 2026]. That matters because enterprise demand for supply-chain intelligence is usually broader than a single material family, and a platform that can carry similar logic across categories has a larger practical addressable market than a point solution tied to one feedstock.
| Market lens | Public evidence | Implication for addressability |
|---|---|---|
| Direct materials procurement analytics | Atmospheric AI focuses on costs, capacity, supply risk, and emissions for raw materials [Axios Pro, November 2026] | Suggests budget overlap with procurement, sourcing, and risk teams rather than only sustainability software |
| Industrial supply-chain modeling | The company builds computational models, described as digital twins, of raw-material production and distribution [Axios Pro, November 2026] [The SaaS News, November 2026] | Points to adjacency with industrial data and decision-support platforms |
| Multi-material expansion | Reported roadmap extends from plastics into chemicals, paper, and metals [The SaaS News, November 2026] [DevCuration, November 2026] | Broadens potential buyer relevance if the same model architecture transfers across materials |
| Sustainability and carbon decision support | Emissions analysis is part of the product brief in funding coverage [Axios Pro, November 2026] [Startup Mag, November 2026] | Creates overlap with carbon accounting and Scope 3 management spend, though the product appears more operational than reporting-led |
The table points to a market that is probably best understood as an overlap zone rather than a clean standalone category. That can be a strength early, because companies often buy into a painful workflow before they buy into a new software label, but it can also make budget ownership and competitive set less tidy.
Regulatory and macro forces also help explain timing, even if the source pack does not cite specific statutes. The presence of emissions and carbon in the product narrative indicates that compliance and decarbonization pressures are becoming part of procurement math, not a separate reporting exercise [Axios Pro, November 2026] [The SaaS News, November 2026]. At the same time, the emphasis on capacity, supply risk, and shipping-route-linked material flows suggests a world in which disruptions and price shifts remain frequent enough to justify more granular external modeling [LinkedIn] [Axios Pro, November 2026].
The substitute market is important here. Large manufacturers already use ERP, procurement suites, supplier risk tools, consultants, and internal category managers to answer versions of these questions, so Atmospheric AI is not creating demand from zero [Startup Mag, November 2026] [Axios Pro, November 2026]. The bet is narrower and more specific: that facility-level computation produces a better answer than conventional supplier questionnaires, historical spend analysis, or broad market indices when the material in question is strategically important.
On public evidence, that thesis has some logic, but the market read remains more directional than measured. There is no cited third-party market-sizing report in the source set, no disclosed pricing, and no public retention data, so the strongest support for category demand today comes from the reported customer set, the named Haleon engagement, and the company’s decision to start in materials where both carbon intensity and sourcing complexity are high [The SaaS News, November 2026] [Startup Mag, November 2026] [Axios Pro, November 2026].
Partially corroborated -- Section relies on named public reporting from Axios Pro, The SaaS News, Startup Mag, DevCuration, the company website, and LinkedIn, but contains no independent third-party market-size dataset for the exact category.
Competition and Substitutes
Competition and Substitutes
MIXED Atmospheric AI appears to sit between enterprise procurement software and industrial intelligence tools, with its differentiation resting less on workflow ownership and more on facility-level modeling of direct materials, especially in petrochemicals and plastics [Axios Pro, November 2026] [Atmospheric AI] [The SaaS News, November 2026].
The public record does not name direct startup competitors, so the clearest way to map the field is by substitute category rather than logo. On one side are incumbent procurement and supply-chain platforms that already sit inside large manufacturers' sourcing processes, but are typically fed by supplier submissions, contract data, and internal ERP records rather than computed plant-level production models [Axios Pro, November 2026]. On another side are specialist sustainability and carbon-accounting tools that help companies quantify emissions exposure, though the reporting here suggests Atmospheric AI is trying to answer a broader sourcing question that joins cost, capacity, risk, and emissions in one decision model [Axios Pro, November 2026] [The SaaS News, November 2026]. A third substitute is still the status quo: consultant-led analysis and supplier relationship management, where category managers piece together fragmented market intelligence manually.
That positioning matters because Atmospheric AI is not claiming to replace all procurement infrastructure. The company says it computes direct materials value chains facility by facility, and Axios describes the product as computational models, effectively digital twins, of raw-material production and distribution [Atmospheric AI] [Axios Pro, November 2026]. If accurate, that places the product upstream of sourcing decisions rather than at the transactional system of record, which can be an advantage early because it narrows the product scope, but it also means incumbents with stronger workflow control could copy adjacent features if the modeled data proves commercially valuable.
The clearest edge visible today is the combination of domain framing and early market access. Atmospheric AI's initial focus on petrochemicals and plastics is narrower than a general procurement suite, but that narrowness may be useful if the company can build hard-to-replicate datasets around plant technologies, efficiencies, energy mixes, and material flows across shipping routes [Axios Pro, November 2026] [LinkedIn, retrieved 2026]. The team profile supports that wedge in part: Roee Goldberg previously co-founded OpenWeb, and founding chief commercial officer Torsten Lichtenau joined after 22 years at Bain, where he was most recently senior partner and global head of carbon transition [Axios Pro, November 2026] [nstartup, November 2026] [Markets Insider, December 2021]. DBL Partners leading the November 2026 seed and adding Vinnie Caruso to the board also gives the company a credible climate-industrial sponsor at an early stage [Axios Pro, November 2026] [tradepoint.io, November 2026].
Still, that edge looks perishable unless the company turns model quality into embedded usage. The reported footprint, more than 50,000 industrial facilities modeled worldwide and an initial group of approximately 100 manufacturers, sounds substantial, but both figures remain lightly corroborated in public sources [Axios Pro, November 2026] [The SaaS News, November 2026] [LinkedIn, retrieved 2026]. If customers treat the platform as a periodic intelligence layer rather than a daily operating surface, switching costs may stay modest, and larger procurement or industrial software vendors could absorb similar analytics into broader suites. Publicly surfaced hiring also suggests the team is still building core product capacity, with two senior full-stack engineering roles in London tied to the Cortex application [freehire.me].
The company looks most exposed where an incumbent already owns buyer attention. Large manufacturers' procurement, sourcing, sustainability, and R&D teams often buy software through established vendor relationships, so a specialist model layer can be strategically useful without becoming budget-critical [Axios Pro, November 2026]. Atmospheric AI is also currently framed around direct materials and an initial petrochemicals wedge, which may limit its relevance outside process-heavy categories unless the expansion into chemicals, paper, and metals lands quickly and with the same model fidelity [The SaaS News, November 2026]. Haleon's reported work with the company on resin sourcing for almost two years is a useful signal, but one named collaboration does not yet show category-wide dominance [Startup Mag, November 2026].
Over the next 18 months, the most plausible competitive scenario is a split market. Atmospheric AI could emerge as a winner if manufacturers decide that supplier-reported data is too incomplete for high-stakes direct-material sourcing and if the company can convert its petrochemical facility models into repeatable procurement workflows across adjacent categories [Axios Pro, November 2026] [The SaaS News, November 2026]. The more likely loser in that scenario is manual, consultant-led sourcing analysis, which is slower to update and harder to scale across thousands of facilities. The reverse case is also straightforward: Atmospheric AI loses ground if incumbent procurement or industrial software vendors add comparable facility-level intelligence before Cortex becomes operationally embedded. In that outcome, the specialist layer risks being treated as a useful input rather than a category owner.
Opportunity
PUBLIC
The prize here is unusually large for a young industrial software company: if Atmospheric AI becomes the system manufacturers trust to price, de-risk, and decarbonize direct materials sourcing facility by facility, it could sit in the decision path for some of the largest cost bases in global industry [Axios Pro, November 2026] [Atmospheric AI].
The headline opportunity is not generic procurement software. It is a higher-stakes position: a system of record and inference for direct materials, where buyers use computed models of production costs, capacity, energy mix, shipping routes, and emissions rather than relying mainly on supplier disclosures or brokered market views [Axios Pro, November 2026] [The SaaS News, November 2026]. That is a reachable outcome, at least in outline, because the company is already described as modeling more than 50,000 industrial facilities worldwide and working with an initial group of about 100 manufacturers, while a named enterprise, Haleon, reportedly worked with it for almost two years on resin sourcing [LinkedIn] [The SaaS News, November 2026] [Startup Mag, November 2026]. Those facts do not prove scale economics yet, but they do suggest Atmospheric AI has moved beyond a concept deck into applied workflows inside a difficult buyer set.
The growth paths are still narrow enough to describe clearly. The table below lays out three plausible ways the company could grow into an outsized outcome, each tied to a specific operating catalyst rather than a general hope that industrial AI spending rises.
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Resin beachhead to multi-material platform | Atmospheric AI starts in petrochemicals and plastics, then expands into chemicals, paper, and metals for the same enterprise accounts | Successful category expansion beyond plastics, which management has already outlined publicly [The SaaS News, November 2026] | The initial focus on plastics gives the company a bounded wedge, while public reporting says the roadmap already extends into adjacent material classes [Axios Pro, November 2026] [The SaaS News, November 2026] |
| Procurement intelligence standard for global manufacturers | Large manufacturers adopt the platform across procurement, sourcing, sustainability, and R&D teams, making it part of recurring sourcing decisions | More named enterprise references like Haleon and broader deployment across the approximately 100 manufacturers already cited [Startup Mag, November 2026] [The SaaS News, November 2026] | The product is positioned for cross-functional use cases, cost, risk, capacity, and carbon, which can widen budget ownership once one team proves value [Axios Pro, November 2026] [freehire.me] |
| Board-backed climate and supply-chain infrastructure play | Atmospheric AI becomes a preferred analytics layer for investors and operators seeking cost and emissions visibility in heavy industry supply chains | DBL Partners' board involvement and commercial build-out under a founding chief commercial officer with deep carbon-transition credentials [tradepoint.io, November 2026] [nstartup, November 2026] | The investor group includes climate and industrial-transition oriented backers, and DBL's Vinnie Caruso is reported to have joined the board, which can matter when selling into long-cycle industrial accounts [Axios Pro, November 2026] [tradepoint.io, November 2026] |
What compounding looks like is straightforward in theory, even if still early in practice. Each deployment can sharpen the company's facility-level models of how materials are produced, what technologies and energy mixes sit underneath them, and how those materials move through shipping routes; if that model quality improves, the product should become more useful for adjacent sourcing categories and for more teams inside the same customer account [LinkedIn] [Axios Pro, November 2026]. Atmospheric AI's positioning also matters here: a system that informs both cost and carbon decisions has a better chance of surviving budget cycles than a single-purpose emissions tool, because it can tie sustainability analysis to procurement economics [Axios Pro, November 2026] [The SaaS News, November 2026]. The early signs of that flywheel are modest but visible in public reporting, namely a two-year Haleon engagement and an initial manufacturer cohort large enough to imply repeated modeling work rather than one-off consulting [Startup Mag, November 2026] [The SaaS News, November 2026].
The size of the win is hard to bracket precisely from public data because no verified market-sizing figure or close public comparable appears in the source set. Still, a reasonable upside frame is that, if the "procurement intelligence standard" scenario plays out, Atmospheric AI could grow into a strategic software asset for industrial and consumer-goods supply chains, valuable either as an independent platform or as an acquisition target for a larger supply-chain, ERP, or industrial-data company (scenario, not a forecast). The evidence for that statement is indirect rather than numerical: the company is addressing direct materials, not back-office spend; it is already selling a cross-functional cost, risk, and carbon proposition; and it has attracted a seed syndicate oriented toward industrial and climate-transition software [Axios Pro, November 2026] [The SaaS News, November 2026] [Startup Mag, November 2026]. In plain terms, if the models become trusted enough to shape sourcing decisions at scale, the company could matter far beyond a typical vertical SaaS vendor.
Partially corroborated -- Relies on named public reporting from Axios Pro, The SaaS News, Startup Mag, nstartup, LinkedIn-sourced profile claims, and company materials. Core opportunity framing is grounded in public product and traction claims, but the most important operating metrics remain lightly corroborated and partly company-driven.
Sources
Open sources
[Atmospheric AI, retrieved 2026] Atmospheric AI | https://atmosphericai.com/
[Crunchbase] Atmospheric AI | Crunchbase Company Profile & Funding | https://www.crunchbase.com/organization/atmospheric-ai
[Axios Pro, November 2026] Exclusive: Atmospheric AI raises $10M to track materials supply chain | https://www.axios.com/pro/climate-deals/2026/11/04/atmospheric-ai-dbl-seed
[The SaaS News, November 2026] Atmospheric AI Raises $10M Seed | https://www.thesaasnews.com/news/atmospheric-ai-raises-10m-seed/
[freehire.me] Atmospheric AI | https://freehire.me/companies/atmospheric-ai
[Markets Insider, December 2021] OpenWeb Partners with NewsGuard to Combat Misinformation Online | https://markets.businessinsider.com/news/stocks/openweb-partners-with-newsguard-to-combat-misinformation-online-1031035921
[LinkedIn, retrieved 2026] Roee Goldberg | Atmospheric AI | LinkedIn | https://www.linkedin.com/in/roeegoldberg/
[DevCuration, November 2026] Atmospheric AI Raises $10.1M Seed for Procurement Models | https://devcuration.com/articles/atmospheric-ai-raises-10-1m-seed-for-procurement
[nstartup, November 2026] Atmospheric AI Raises £7M to Map Industrial Supply Chains | https://nstartup.me/news/4e73b8c5-4a3d-4ff9-affd-54a70d8fa04c
[Startup Mag, November 2026] Atmospheric AI raises £7m from DBL Partners and RAISE | https://www.startupmag.co.uk/funding/atmospheric-ai-2026-seed-funding/
[Funding Spotter, August 2026] Atmospheric AI Funding | https://fundingspotter.com/
[Gartner Peer Insights, 2026] Gartner Peer Insights | https://www.gartner.com/reviews/
[G2, 2026] G2 | https://www.g2.com/
[GeniusFirms, October 2026] GeniusFirms | https://geniusfirms.com/
[LinkedIn] Roee Goldberg | Atmospheric AI | LinkedIn | https://www.linkedin.com/in/roeegoldberg/
[tradepoint.io, November 2026] Atmospheric AI secures $10M seed for procurement modeling push | Unite.AI | https://tradepoint.io/atmospheric-ai-secures-10m-seed-for-procurement-modeling-push-unite-ai/
Articles about Atmospheric AI
- Atmospheric AI Models 50,000 Factories to Price Carbon Into the Supply Chain — The London startup's platform, Cortex, computes cost and emissions for raw materials, starting with plastics. Its seed round was led by DBL Partners.