Atmospheric AI Raises $10.1M Seed for Procurement Models

Asking a supplier to explain a price brings the factory behind the quote into the conversation. Atmospheric AI is building software for that exchange, giving manufacturers a way to examine production costs, emissions and supply exposure several steps upstream of their direct vendor.

The London-based company’s $10.1M Seed round, led by DBL Partners, will support expansion beyond plastics into chemicals, paper and metals. Its November 5, 2026 announcement describes the raise as $10M, while the company’s site banner and participating investor RAISE Ventures report the more precise $10.1M figure. The commercial proposition is unusually concrete: help the person agreeing a materials price understand the industrial assumptions inside it.

A Seed Round With a Procurement Assignment

Co-founders Roee Goldberg, CEO, and Toby Darbyshire are developing Atmospheric AI around the information available to sourcing teams. DBL’s investment brings Vinnie Caruso onto the board, while RAISE Ventures identifies returning investors including Broadscale, tiny.vc and Anamcara. Norrsken Evolve has also publicly confirmed follow-on participation.

The financing connects technical coverage with commercial reach. Atmospheric AI says the proceeds will help it work more deeply with an initial group of around 100 manufacturers as it adds material categories. That figure describes the company’s stated collaboration plans; the announcement supplies no audited customer count, revenue figure or valuation.

For a procurement buyer, the useful question is what those additional categories make possible at the purchasing desk. A sourcing team can compare a new quote with an old one and still lack an explanation for the difference. Atmospheric AI is seeking a place in the work required to construct that explanation, where a technical estimate has to become evidence somebody can use with a supplier.

Why a Facility Model Changes the Discussion

Atmospheric AI starts with chemical engineering, plant economics and the processes that turn feedstocks into materials. The company says it operates around 50,000 facility models connected across production and logistics networks. Its approach aims to estimate cost, carbon and exposure without depending on access to suppliers’ internal operating data.

Those estimates have a different job from the price on a commercial offer. A supplier’s price can reflect its negotiating position, contract terms and willingness to sell as well as the cost of production. An independently computed cost gives the buyer another reference point. It helps the buyer identify which part of an explanation deserves a closer look, without making the supplier’s commercial decision disappear.

That distinction also gives the software an accountability problem of its own. A model used in a negotiation must be explainable enough for the buyer to understand its assumptions and boundaries. Procurement professionals remain responsible for choosing a source and agreeing terms; the analytical value depends on whether they can connect an estimate to the material, facility and purchasing situation in front of them.

Customer Evidence Stays Close to Its Source

Haleon appears in the company’s announcement as a manufacturer working with Atmospheric AI on resin sourcing. A testimonial attributed to Chris Low, Haleon’s Head of Sustainability Procurement, describes using estimated cost, carbon footprint and supply-chain risk as inputs to sourcing. This is a named commercial relationship supported by a company-published account, with no independently audited savings figure disclosed.

A buyer considering the platform would still need to establish how it performs on that buyer’s own materials. The relevant comparison could include the quality of the assumptions, how often an estimate changes, and whether the sourcing team can explain its use internally. The industrial procurement work already attracting capital makes this a useful distinction for buyers: an analytical input has to earn its place in a purchasing workflow. A model can sharpen the questions asked before a contract is agreed.

The relationship also helps explain why this category requires more than a polished interface. Material expertise and purchasing judgment meet inside the same decision. A sourcing professional needs to understand which conditions would change an estimate, while the software provider needs to learn which uncertainties actually alter a procurement choice. That exchange creates an ongoing deployment job alongside the modeling work.

Expanding the Materials and the Commercial Team

Rob McInerney identifies himself as Atmospheric AI’s Founding CTO. Dr Torsten Lichtenau is its Founding Chief Commercial Officer, joining after 22 years at Bain & Company, including senior-partner and global carbon-transition leadership roles. Their responsibilities put technical development and enterprise adoption alongside each other as the company broadens its scope.

Chemicals, paper and metals bring different production processes and purchasing requirements. Expanding into them means deciding where engineering detail materially improves a buyer’s understanding, then making that detail usable during sourcing. The financing supports that ambition; it provides no basis for assuming identical results across every material or customer.

Atmospheric AI’s current careers page lists modeling, engineering and customer-deployment roles, with a remote-first structure and London hub. The hiring mix is consistent with a company that has to build the industrial model and help customers use it. As its coverage expands, those teams will keep meeting at the same practical point: a manufacturer needs to agree a price, and wants a defensible account of the plants and supply routes behind it.