Manufacturing AI platform CADDi has closed a $114 million Series D at a $1.2 billion valuation, giving the Tokyo-based company fresh firepower to convert its Japanese factory-floor foothold into a full-scale North American expansion. The round, announced on September 15, 2026, brings the company's total funding to roughly ¥36.3 billion ($245 million) since it was founded in 2017.
What CADDi actually sells
CADDi's software structures the messy trove of drawings, specifications, quotes and production history that never made it into a modern ERP. Its CADDi Drawer product lets buyers and engineers search decades of legacy CAD files as if they were LLM-readable documents, and the newer CADDi Explorer, CADDi Agent and factory workflow tools chain that data into procurement, quality and supplier decisions. CEO Yushiro Kato told Fortune the pitch is simple: "I've never seen anybody who uses LLMs to do design reviews because it doesn't understand drawings or CAD."
Who's writing the checks
The Series D was joined by Moore Strategic Ventures, Coreline Ventures, Salesforce Ventures, Toyota Motor's Woven Capital, Recruit Holdings' HR Tech Fund, Atomico, Globis Capital Partners and JPS Growth Investment. The syndicate reads like a strategic bet on manufacturing intelligence: two of Japan's most influential industrial groups (Toyota and Recruit) sit next to enterprise software backers, which lines up with CADDi's claim that more than half of the Nikkei 225's manufacturers are already customers.
Where the money goes
The proceeds are earmarked for North American growth from CADDi's Chicago office, deeper AI R&D on its manufacturing-specific models, and hiring across product and go-to-market. Kato has said the U.S. rollout is aimed at large industrial buyers and their supply chains, where the same fragmented CAD problem exists at scale. The deal also lands during a broader wave of physical-AI investment that has already produced 2026 rounds for Noetive, TypeSafe AI and Vecna Robotics.
Why it matters
Manufacturing has been slow to adopt LLMs because the raw material — drawings, PLM data, machining specs — isn't text a general-purpose model can reason about. CADDi is trying to be the layer that changes that, and the Series D price tag suggests investors think that layer is worth staking now, before hyperscaler-scale rivals or larger PLM vendors get there.
Reporting based on coverage from SiliconANGLE, Fortune, and TechStartups.