- Partly has raised $50 million in Series B funding led by DST Global at a $500 million valuation.
- Its Interpreter model scored 60% F1 on the company’s own benchmarks, compared with 1–5% for general AI without domain data.
- DST Global, backer of Facebook, Alibaba, and Airbnb, is making an unusual move into industrial-vertical AI.
Ask GPT-5 to identify the exact parts needed to repair a 2019 Toyota Camry after a rear-end collision. According to one startup, it will get the answer right about 5% of the time. That startup just convinced DST Global, the firm behind some of the most valuable consumer internet companies in history, to lead its $50 million Series B.
Partly, founded in Christchurch, New Zealand in 2020 by Levi Fawcett, ex-engineer at Rocket Lab and four-time founder, alongside co-founders Nathan Taylor and Taylor Howes, announced the raise at a $500 million valuation.
The company, now headquartered in Austin, Texas, with a second office in San Francisco, employs 160 people across 20-plus countries.
“Not since the creation of the assembly line or EVs has the auto industry experienced significant innovation that simultaneously improves operational efficiency, industry profitability, and consumer value,” Fawcett says.
Why general AI gets this wrong
Partly builds Interpreter, which it describes as the only foundation model purpose-built for automotive parts. The problem it solves is called fitment: identifying the exact correct parts for a specific vehicle, accounting for manufacturer variations, trim levels, build-plant differences, and model-year changes across hundreds of millions of configurations. General models were not built for this.
“If you ask GPT-5 or Claude Opus 4.8 a question about parts or automotive, they’re going to be very bad — they’ll get it right about 5% of the time. On complex jobs, Interpreter is about four times more accurate than a human, and an order of magnitude more accurate than those models,” Fawcett tells Tech Funding News.
Partly has published its own benchmark to support those claims. Tested across 50 Toyota repair jobs, Interpreter achieved 60% F1, a combined measure of precision and recall, against 1–5% F1 for general AI models without catalogue data. The company was trained on five years of human feedback and synthetic data, plus agreements with more than fifty manufacturers.
Incumbents, including Solera, Mitchell1, and ALLDATA, offer workflow software and reference databases built on legacy infrastructure. None operates at the foundation model layer. That distinction is Partly’s core claim to the market.
What DST saw
DST Global’s track record includes Facebook, Alibaba, Airbnb, Spotify, and Anthropic, all of which are large-ticket bets on category winners in the consumer internet. An industrial B2B company serving collision repair shops is a different profile.
Fawcett says the underlying logic holds. “DST have been making a lot of bets around physical world AI — bringing AI into the physical world. We’re assisting technicians, helping them do better work, taking away the admin,” he tells TFN.
The firm co-led OpenEvidence’s $250 million Series D in January at a $12 billion valuation and participated in Waymo’s latest round — both plays on AI entering physical operations.
Partly fielded interest from Andreessen Horowitz and Iconiq before DST came in. The market scale drove the decision. “You can build a trillion-dollar business just within this domain,” Fawcett adds.
The $50 million is a partial close. Fawcett does not expect to raise additional capital until next year. The funds will go towards computing for training version seven of Interpreter, team expansion across Europe, and US sales and marketing. It brings total funding to $92.4 million, following a $37 million Series A in December 2022 led by Octopus Ventures at a $180 million valuation, with participation from Square Peg, Blackbird, Shasta Ventures, and Ten13.
The US collision repair market exceeds $100 billion annually, spread across roughly 250,000 independent shops. Fawcett believes the adoption window has opened.
“It really has started working in the last six to nine months. When you talk to the CEOs at any of the large repairers, that is the number one thing they are talking about,” he concludes.
DST has seen this pattern before: a large, fragmented, seemingly unglamorous market sitting one technology cycle behind where it should be. It backed the infrastructure play each time.
Whether automotive repair follows the same curve as the markets DST bet on before will determine if this is another prescient early call or the moment DST found the edge of its own pattern.
