Atomic Raises $12.5 Million to Let AI Agents Place Purchase Orders Themselves
A Boston startup that once helped Tesla navigate its Model 3 production crisis has raised fresh capital to automate one of the least glamorous but most consequential activities in modern commerce: deciding what to buy, when to buy it, and hitting "send" on the order.
Atomic announced Friday it closed a $12.5 million Series A co-led by Klass Capital and Madrona, with participation from DVx Ventures, Alumni Ventures and Sandberg Bernthal Venture Partners. The round brings total disclosed funding to just over $15 million and arrives as the company claims it now autonomously handles 90% of purchasing decisions across hundreds of DoorDash's DashMart convenience sites.
That last detail is the headline. For decades, supply chain software has whispered suggestions to procurement managers, who then logged into separate systems to place orders. Atomic's pitch is more aggressive: its platform doesn't just recommend purchase orders. It executes them, with human buyers stepping in only for exceptions or anomalies.
Jon McNeill, founder of DVx Ventures and a former president at Tesla, compares the shift to autopilot in cars. "DoorDash is running, I think, 90% of its purchasing across hundreds of sites," he told TechCrunch. McNeill says Atomic's annual recurring revenue has quintupled since the beginning of 2026, though the company declined to disclose specific figures.
The three co-founders brought unusual pedigrees to the problem. CEO Michael Rossiter, Chief Product Officer Neal Suidan, and CTO Jeff Goodrich all led sales and operations planning at Tesla during the chaotic 2017-2018 Model 3 production ramp, eventually building a 50-person planning engineering team there. Atomic was incubated at DVx Ventures, McNeill's venture studio, and raised a $3 million seed round led by DVx in April 2024.
DoorDash's deployment offers a window into what autonomous purchasing looks like in practice. The company's DashMart fulfillment network migrated to Atomic in roughly three months with what Atomic describes as "feature parity and zero business disruption." After going live, DashMart's team built new supplier logic in about an hour; Atomic applied the rules daily, shifting more volume to primary suppliers and improving gross margins, according to the company. "Atomic has helped us plan and manage inventory more proactively," Nick Palefsky, director of new business verticals at DoorDash, said in a statement.
The platform sits alongside existing ERP systems from NetSuite, SAP, Microsoft Dynamics and Oracle, rather than replacing them. Atomic's sales-and-operations-planning layer can be deployed in about 30 days using existing data, while full platform deployment across purchasing, production and allocation typically takes four to six months. The software includes Nucleus, a set of AI agents that handle S&OP prep, inventory questions and supply-risk checks, then execute routine purchase orders while humans manage exceptions.
Good Chop, a meal-delivery brand owned by HelloFresh, reduced inventory on hand from eight or nine weeks down to four weeks while more than doubling revenue using Atomic as what the company calls its "core engine" to plan supply and place purchase orders. "Atomic is the core engine we use to plan supply," Seb Tron, president of Good Chop, said. Other named customers include skincare brand Starface, footwear maker OOFOS, electrolyte drink company LMNT and watch brand Vincero.
Atomic says it plans to extend its platform from planning and decision support into what it calls a "control system" that links business objectives to day-to-day decisions. Madrona describes the approach as a "Reasoning Flywheel," where decisions compound over time and gradually shift from recommendations to autonomous action by AI agents. "We call that compounding loop the Reasoning Flywheel," Matt McIlwain, managing director at Madrona, wrote in a post explaining the firm's investment. Adrian Schauer of Klass Capital and McIlwain joined Atomic's board as part of the Series A.
The company is hiring forward-deployed engineers and senior software engineers across New York, Boston, Austin and Reno, with remote options available.
Atomic was the third AI purchasing startup to raise a Series A in two weeks, according to industry publication The Conveyor. Magentic raised an $18 million round led by Felicis on September 21, and Montreal-based Axya closed C$17 million led by McRock the same day Atomic announced its round. The flurry of activity comes as established supply chain planning vendors like Kinaxis, o9 Solutions and SAP IBP command multi-billion-dollar valuations. Kinaxis trades at a market cap of roughly CAD $4.85 billion; o9 Solutions reached a $3.7 billion valuation in 2023; and Panasonic acquired Blue Yonder for an $8.5 billion enterprise value in 2021.
Rossiter told Axios Pro the company has already received inbound investor interest and could raise a Series B within the next year. Whether that timeline holds may depend on how comfortable more buyers become with letting software place orders on its own, a leap of faith that requires not just technical sophistication but a tolerance for the occasional algorithm-driven mistake.
