AI Under the Hood: Partly's $500M Plan to Rebuild Auto Repair

AUSTIN, Texas – June 23, 2026 – The automotive repair industry, a sprawling, $100 billion-plus sector in the U.S. alone, has long operated on a foundation of experience, intuition, and sprawling, often-incompatible paper catalogs. It's a world of grease-stained hands and frustrating phone calls to find a single, correct part. But a seismic shift is underway, powered not by horsepower, but by processing power.

Partly, an AI infrastructure company, today announced it has closed a $50 million Series B funding round, catapulting its valuation to $500 million. The investment, led by the formidable DST Global Partners—a firm known for backing generational companies like Meta, Spotify, and AI pioneer Anthropic—coincides with Partly's immediate and ambitious expansion into the U.S. market. The company is betting that its specialized AI, dubbed 'Interpreter', can become the new digital nervous system for an industry desperately in need of a modern overhaul.

Decoding the 'Interpreter': AI for the Analog Age

At the heart of Partly's strategy is a problem every mechanic and DIY enthusiast knows intimately: fitment. Identifying the exact correct part for a specific vehicle is a Herculean task. A single car model can have thousands of variations based on its model year, trim level, build plant, and mid-cycle updates, resulting in hundreds of millions of possible configurations across the automotive landscape. This complexity is where generalist AI models falter.

According to the company's internal benchmarks, large language models like GPT-5 or Claude Opus 4.8 are highly inaccurate, correctly identifying complex parts only about 5% of the time. This is the challenge Partly has spent five years and over $10 million in research to solve. The result is 'Interpreter', which the company describes as the world's only foundation model purpose-built for automotive parts. Unlike its generalist cousins, Interpreter is a specialist. It has been meticulously trained on a diet of technical diagrams, damage photos, manufacturer agreements with over 50 automakers, and continuous live data from repair workflows.

The model is multimodal, capable of ingesting a photo of a damaged car, a technician's voice note, or a complex schematic and translating it into a standardized, actionable list of required parts. Its performance is striking. In one benchmark test involving 50 Toyota repair jobs, Interpreter achieved an F1 score—a statistical measure of accuracy—of 60%, a full order of magnitude better than general AI. The company claims its model is about four times more accurate than a human expert on complex jobs. By creating a foundational infrastructure layer, Partly isn't just building an app; it's building the underlying language that could one day power all transactions in the parts supply chain.

The $50 Million Bet on a $100 Billion Problem

The backing of DST Global Partners is more than just a financial boost; it's a powerful signal. The venture firm has a legendary track record of identifying category-defining internet companies. Its recent interest has expanded into what some analysts call "physical world AI," targeting companies that use sophisticated algorithms to solve complex industrial problems. This investment places Partly alongside other DST-backed ventures like Waymo, which are applying AI to the physical realities of the automotive world.

This capital is being deployed to tackle a well-documented and costly problem. The American automotive repair supply chain bleeds billions of dollars annually from inefficiencies. A recent survey found that nearly half of all repair shops cite delays in parts delivery as the primary reason for longer service times. This isn't just an inconvenience; it leads to lost revenue, strained customer relationships, and operational gridlock. Shops using early versions of Partly's system have reported processing orders nine times faster while reducing incorrect part returns by a factor of 2.4—a direct injection of efficiency into their bottom line.

"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," said Levi Fawcett, CEO and Co-founder of Partly. "We have spent five years building the AI infrastructure layer that the industry has been missing. The model architecture is extremely nuanced, there's a reason general models don't solve it, and why we've been able to own the frontier AI here."

Austin's New Frontier: From Silicon Hills to Engine Blocks

To spearhead its American offensive, Partly has anchored its U.S. operations in Austin, Texas. In a significant commitment, the company has relocated its core executive team, including CEO Levi Fawcett, to the city. It is now actively recruiting for high-skilled roles in engineering, business development, and product management to serve the nation's approximately 250,000 repairers.

The move cements Austin's status as a burgeoning tech hub that is diversifying beyond consumer software into the more complex world of industrial AI. While the U.S. market has established players like Solera and Mitchell1 offering workflow software and reference databases, Partly aims to differentiate itself by operating at a more fundamental level. It isn't selling another piece of software as much as it's offering access to an intelligent utility, a new form of digital plumbing for the entire industry. The market appears ready, with industry reports showing that while only a fifth of aftermarket businesses have deployed enterprise AI, another 20% are actively planning to, indicating a sector at a technological tipping point.

Beyond the Wrench: AI as the Next Assembly Line

Fawcett's comparison to the assembly line is telling. That innovation wasn't just about making one factory faster; it restructured the entire logic of manufacturing. Similarly, the long-term vision for an AI like Interpreter extends far beyond the local repair shop. A universal, intelligent parts language could streamline everything from insurance claim processing and automated damage estimates to providing real-time feedback to manufacturers about component failures in the field.

For the thousands of independent shops that form the backbone of the industry, this technology represents a potential leveling of the playing field. Instead of being a disruptive threat, it's being positioned as an empowering tool—a co-pilot that minimizes costly errors, frees up technicians to do the work they are trained for, and ultimately helps them compete in an increasingly complex market. The promise is a future where the right part is identified in seconds, not hours, and a vehicle's repair time is dictated by the mechanic's skill, not the supplier's logistical hurdles. With this expansion, the future of auto repair is increasingly being written not in oil and steel, but in data and algorithms.