CADDi, a Japanese startup that uses AI to help large global manufacturing companies optimize their supply chains, has raised $38 million in new funding from U.K.-based venture capital firm Atomico.
The new funding values CADDi at $470 million, the company said. The investment round, which the company is classifying as a “Series C extension,” brings the total amount of venture funding CADDi has raised since its founding in 2017 to $202 million.
CADDi had announced an $89 million Series C round in July 2023, with DCM Ventures, Globis Capital Partners, Minerva Growth Partners, and WiL (World Innovation Lab) all participating in the round.
CADDI, which has headquarters in both Tokyo and Chicago, already has some U.S. customers and is planning to use the new funding to ramp up its U.S. expansion. The company also plans to double the number of software engineers it employs, from 150 to 300, Yushiro Kato, the ex-McKinsey consultant who is CADDi’s cofounder and CEO, said. The company currently employs 600 people in total.
CADDi sells software that addresses a problem many large manufacturing companies have: They have too many similar parts being provided by too many different suppliers. CADDi ingests technical drawings of a part and then searches the company’s own data to find similar components—or, in some cases, identical parts—that are already in the company’s inventory or that it has bought previously. It also allows employees to search for parts using keywords that may be used in component descriptions.
Manufacturers can use CADDi’s software to avoid component duplication, optimize supply chains for parts, such as fasteners, that may be common to many different products, and potentially reduce the number of suppliers they are using. That in turn can save costs by reducing the amount of time it takes to source a part and avoiding duplicative procurement processes and related paperwork. It also potentially lets the manufacturer get better prices on parts by purchasing higher volumes from a smaller number of suppliers.
Kato told Fortune that the startup’s customers are primarily companies that make machinery for factories—for instance, food production machinery, packaging machinery, and semiconductor manufacturing machinery—and also automotive and auto parts companies.
He said that one automotive parts customer reduced the number of different fastener SKUs it was using by 60% thanks to CADDi’s software.
Automotive company Subaru said in a statement provided to Fortune that using CADDi’s software had saved it “hundreds of hours per month” in the time employees spend searching for technical drawings.
DENSO, the Japanese auto parts company, has a partnership with CADDi and said in a statement that the company’s software allows younger, less experienced workers source components faster. Previously, procurement processes were dependent on the knowledge of veteran workers, many of whom are now approaching retirement age, DENSO said. It also said that it was working with CADDi to develop additional product features, such as the ability to search three-dimensional drawings, as well as two-dimensional engineering schematics.
Kato declined to reveal the company’s current revenues or the total number of customers currently using the platform. But he said the company was targeting $1 billion in revenue from its software platform by 2030.
When CADDi was first founded, it functioned as a kind of “Amazon marketplace for machinery components,” Kato said. Customers would send it technical drawings or engineering specs for parts that they needed, and CADDi would go out and source those parts for the customer, acting as a kind of parts broker. In order to do this efficiently though, the company wound up developing a lot of its own software, including AI models that can do searches based on technical drawings. Three years ago, Kato and his cofounder Aki Kobashi, who is CADDi’s chief technology officer, pivoted away from being a parts marketplace, instead selling the AI software it had developed as a cloud-based platform to manufacturing companies.
