If there’s one trend in the auto parts industry that’s been making my team and I jump out of our chairs over the past five years, it’s the rise of autonomous vehicles (AVs). Back in 2018, most of our work was focused on traditional internal combustion engine (ICE) parts—door handles, radiator grilles, dashboard trim pieces that had been around for decades. Now, 70% of our projects are for AV-specific components, and that shift has forced us to completely rewrite how we design and build auto parts moulds. As a third-generation owner of our family-run mould shop, I’ve always believed that the best moulds come from listening to both our customers and the end product, and AVs are teaching us lessons we never could have learned from ICE cars. Today, I want to walk you through what it actually takes to develop auto parts moulds for autonomous vehicles, from the first blueprint to the final test run, and why it’s not just about precision—it’s about thinking like a self-driving car. Auto Parts Mould
Let’s start with the foundation: AVs don’t just look different from traditional cars—their parts have to perform differently. Take a close look at a Tesla Model 3’s interior or a Waymo driverless taxi’s dashboard, and you’ll notice there are fewer physical buttons, more seamless surfaces, and components that work in tandem with the car’s sensors. Every mould we build has to account for these unique demands, starting with a collaborative design phase that’s way more hands-on than our old process. Five years ago, we’d get a 2D CAD drawing from an auto manufacturer, tweak it for manufacturability, and send it back. Now, we sit in on weekly video calls with their engineering, software, and quality teams for 8 to 12 weeks before we even touch a mill. Last year, we worked with a major German AV startup on a mould for their side mirror housing, which has to house a 360-degree camera, a tiny LiDAR sensor, and a tiny motor that adjusts the mirror angle. Their software team told us the housing can’t have any air bubbles larger than a grain of sand because even a tiny imperfection would scatter the sensor’s light and throw off the car’s object detection. That level of detail wasn’t on our radar for traditional mirror moulds—those just needed to look good and hold the glass.
The biggest mistake new mould makers make when switching to AV parts is treating them like traditional auto parts. AV parts have dual lives: they have to be durable enough to survive 15 years of road vibration, temperature swings from -40°F to 180°F, and exposure to UV light, but they also have to be precise enough to align with sensors that can detect a 2-inch object at 100 meters. That’s why our design process now includes a “sensor alignment check” step that we never used before. For the LiDAR housing project I mentioned earlier, we used 3D printing to make a test mockup of the mould, then 3D-printed a prototype part out of the same polycarbonate blend the customer specified. We then mounted a small LiDAR unit to it and ran tests in our in-house test lab, bouncing light off the part surface to make sure there were no micro-textures or warps that would distort the signal. It took three rounds of tweaks to the mould’s core and cavity to get the part within 0.02mm of tolerance—something that would have been overkill for a regular mirror, but non-negotiable for AV.
Once the design is locked in, the next step is material selection, and this is where AV moulds differ drastically from our old work. For ICE parts, we used common plastics like ABS or polypropylene, which are cheap and easy to mould, as long as they meet basic durability standards. For AV parts, we’re working with specialized engineered plastics, fiber-reinforced composites, and even some transparent ceramics for sensor covers. But here’s the catch: the mould has to be built to handle these exotic materials. Last year, we took on a project for a Canadian AV maker that needed a mould for their windshield sensor shroud, which has to be transparent to infrared light and resistant to bird strikes and road debris. The material they chose was a polycarbonate blend with 20% glass fiber, which is strong but abrasive—way harder on mould steel than regular plastic. We used P20 tool steel, which has a higher chromium content than our standard S7 tool steel, and we added a polished surface coating that reduces friction by 30% to extend the mould’s life. We also had to adjust the mould’s cooling system: glass fiber materials generate more heat during injection, so we added 12 extra cooling lines around the part’s edges to prevent warping. If we’d used our old mould build for this project, the mould would have worn out after 10,000 parts, and half the sensor shrouds would have been misaligned for the car’s camera.
Another key factor in AV mould development is precision and repeatability. Traditional auto parts can have tiny variations—most manufacturing tolerances for door handles are around 0.1mm, which is invisible to the naked eye. For AV parts, that tolerance drops to 0.02mm, and even smaller for sensor-related components. That means our mould-making process has to be far more controlled than it was five years ago. We now use a coordinate measuring machine (CMM) to check every single feature of the mould core and cavity before assembly, and we run a 24-hour “hold test” where we leave the mould assembled at 70°F and 50% humidity for a full day to make sure it doesn’t shift. For a recent mould for an AV’s center console, which has a touchscreen interface integrated into it, we had to ensure that the mounting points for the screen were within 0.015mm of tolerance. If those points were even 0.02mm off, the screen would sit at a slight angle, and the car’s operating system would have a hard time calibrating the touch input. We caught that issue during the hold test—we noticed a small warp in the mould’s core that we’d missed during the CMM check—and made a tweak to the cooling lines that fixed it. If we’d shipped that mould, the customer would have had to scrap thousands of parts once they started assembly, which would have cost them millions of dollars in delays.
Testing is another stage that’s completely different for AV moulds. In the past, we’d run a mould for a few hundred parts, check them for surface imperfections, and ship it to the customer. Now, we run a “production test” where we make 10,000 parts in our own in-house facility, then subject them to a battery of tests that go way beyond what traditional parts need. For the LiDAR housing mould, we ran vibration tests that mimicked 15 years of road travel, temperature cycling between -40°F and 180°F, and UV exposure equivalent to 10 years in Arizona. We also sent parts to a third-party lab to test their ability to block electromagnetic interference (EMI), because if a LiDAR housing absorbs too much EMI, it can interfere with the car’s radar and create a blind spot. That test took three months to complete, and we had to adjust the mould’s wall thickness by 0.5mm to add a thin layer of EMI shielding that the customer had missed in their initial design.
We’ve also learned that sustainability is becoming a non-negotiable for AV moulds—and for the auto parts industry as a whole. EVs and AVs are marketed as greener, so their parts have to match that reputation. That means we’re building moulds that use recycled materials, run on less energy, and produce less waste. For a recent project for a U.S.-based AV startup, we designed a mould that uses 30% post-consumer recycled polypropylene for the interior door panels, instead of virgin plastic. We adjusted the mould’s venting system to avoid air bubbles in the recycled material, which has more inconsistent molecular structure than virgin plastic. We also switched our mould-making process from gas machining to waterjet cutting, which uses 50% less energy and produces no hazardous waste. The customer loved it—they were able to market their AV as using 40% recycled interior parts, which helped them win a contract with a major ride-sharing company.
Of course, developing AV moulds isn’t without its challenges. The biggest one right now is that the AV industry is still changing fast—new sensor technologies are being developed every six months, and regulatory standards for AV parts are still being written across different countries. Last year, we had a project for a European AV maker that required a mould for a camera housing that had to meet new European Union standards for sensor alignment, which weren’t even finalized when we started the project. We had to adjust the mould’s design three times over 12 months to meet the final standards, which added 20% to the project’s cost and delayed delivery by four weeks. That’s why we’ve built a small team of regulatory experts in-house who track changes in AV standards across the U.S., Europe, and Asia, so we can adjust our designs quickly without derailing projects.
Another challenge is the cost: AV moulds are way more expensive to build than traditional moulds, often costing 2 to 3 times more, because of the precision, specialized materials, and testing required. But the payoff is huge—AV manufacturers are ordering 50,000 to 200,000 parts per model, compared to 10,000 to 50,000 for traditional ICE models, so the return on investment for high-quality AV moulds is much faster. We’ve seen that firsthand: our first AV mould project in 2019 was for a side mirror housing, and the customer is still ordering parts from that mould four years later, with no major issues. That reliability is why we’ve built long-term relationships with 12 of the top 15 AV startups in North America and Europe.
At the end of the day, developing auto parts moulds for autonomous vehicles is less about building a better mould, and more about building a mould that works for a car that thinks for itself. Every decision we make—from the tool steel we choose to the cooling lines we add to the testing we run—has to be rooted in the needs of the sensors, software, and safety systems that make an AV work. It’s been a steep learning curve, but as someone who’s been in this business for 35 years, I’ve never been more excited about what’s ahead. The AV revolution isn’t just about new cars—it’s about reimagining how we make every part that goes into them, and that’s a challenge we’re ready to take on.
Auto Parts Mould If you’re an AV manufacturer, a ride-sharing company, or an auto parts brand looking for reliable, precision-built moulds that meet strict safety and performance standards, we’re here to help. Our team of mould designers, engineers, and testers has years of experience working with AV-specific components, and we can tailor a mould solution to fit your project’s timeline, budget, and regulatory needs. Don’t let unreliable moulds derail your next AV launch—reach out to our team today to discuss your requirements and get a custom quote.
References
- Society of Automotive Engineers (SAE). (2022). Mould Design and Manufacturing for Autonomous Vehicle Components. SAE International.
- Automotive Industry Action Group (AIAG). (2021). Quality Standards for Advanced Driver Assistance Systems (ADAS) Part Moulding. AIAG.
- International Organization for Standardization (ISO). (2023). ISO 19085: Road Vehicles – Moulded Components for Autonomous Vehicles – Requirements and Test Methods. ISO.
- Automotive Manufacturing Solutions. (2023). "Precision Moulding: The Backbone of Autonomous Vehicle Sensors." Automotive Manufacturing Solutions.
- Deloitte. (2022). Global Automotive Moulding Market Report: Trends and Opportunities for AV Components. Deloitte Insights.
Taizhou Tangwei Mould Co., Ltd.
Address: Mould City, Xicheng, Huangyan, Taizhou, Zhejiang, China
E-mail: cjc@twmould.com
WebSite: https://www.tangweimould.com/
