The Role of Open Source in Automotive Chiplets: A Vision for a Modular, Sustainable Future
The automotive industry is standing at a crossroads. As vehicle complexity increases and silicon costs skyrocket, traditional approaches to chip design are becoming unsustainable. The cost savings from ECU consolidation and zonal architecture transformation work is being offset by the increasing costs of the few, yet very complex and expensive chips needed by the central compute function. Enter: chiplets— a way to distribute chip complexity to allow usage of cheaper components.
Practically all Automotive OEMs and Tier1s are working with Semiconductor companies towards enabling chiplets in Automotive. Given the significant collaboration required by many players to make chiplet vision true, a few Automotive chiplet ecosystems have formed over the last 2 years. ChipFlow, for instance, has recently joined the Automotive Chiplet Forum, an initiative led by Europe’s imec but with a seemingly global membership footprint. With all the progress the ecosystems have made, there seems to be a problem - the time and associated cost it takes for the chiplet ecosystems to make them a reality in production cars.
To build a commercially viable design ecosystem around chiplets, especially in a safety critical environment as Automotive, is a large undertaking requiring significant upfront investments. The Automotive industry is notorious for wanting to push the innovation cost burden to its suppliers and EDA / IP / Silicon companies are hesitant to push fast forward until there are clearer standards and volume predictions available. In the current way the semiconductor industry is wired, it is the buy-in and progress by the EDA / IP / Silicon companies that is required before the chiplet design solutions will start to further accelerate.
At ChipFlow, we believe there’s a better way forward. Founded by Rob Taylor and Tomi Rantakari, long-time leaders in the automotive open source movement, our mission is to bring the power and flexibility of open source to the world of automotive chiplets. The parallels with the software world are undeniable, and we’re betting on history repeating itself, only this time it’s hardware.
The Open Source Playbook: From Software to Silicon
It wasn't long ago that open source software was considered too risky or niche for the automotive industry. Fast forward to today: over 70% of the automotive software stack is open source, a monumental shift from 0% just 15 years ago. Initiatives like GENIVI (now COVESA) paved the way, creating common middleware platforms that all OEMs could build upon, leaving proprietary development only for areas that actually made competitive sense..
Automotive companies have learned to collaborate on the infrastructure, and compete on the experience. Open source has become a strategic tool, not just a cost-saving mechanism, but a path to resilience, interoperability, and innovation.
Chiplets: The Next Frontier
Now, we’re seeing the same dynamic unfold in silicon. As the industry grapples with skyrocketing BoMs, single-source dependencies, and design inflexibility, chiplets offer a compelling alternative: a modular approach that allows companies to integrate best-of-breed IP, much like software components.
But to make this work, a new model is needed. One that balances commodity and differentiation. At ChipFlow, our vision is a "pick and mix" approach to chiplet design:
Commodity IP blocks (like USB, PCIe, CAN, etc.) that are common across all OEMs should be open source. Frankly there’s simply no competitive advantage in each OEM reinventing these same building blocks.
Proprietary IP should be focused where it matters: performance, safety features, AI acceleration, brand-specific features. Effectively all of the areas that define true product differentiation.
Learning from Deepseek: The Open Source Compounding Effect
What’s happening in hardware today echoes trends in AI. Projects like Deepseek demonstrate the compounding value of open source: once a foundation is laid and widely adopted, innovation accelerates across the entire ecosystem.
Contributors build on top of each other’s work, reduce duplication, and focus their efforts where they really matter.
Similarly, in automotive silicon, once a core library of verified, production-grade open source IP exists, the ecosystem can begin to compound. OEMs and Tier 1s can mix and match components, focus their ASIC spend more strategically, and reduce both cost and time to market.
A European Imperative
This shift is particularly important for Europe, home to a dense network of legacy automotive OEMs and Tier 1s. These companies are facing intense global competition and increasing pressure to localize supply chains. The EU Chips Act presents a unique opportunity to invest in open source chiplet IP as a shared infrastructure—fostering a homegrown, sustainable silicon ecosystem that can serve multiple players without duplicating effort.
ChipFlow is already taking steps in this direction, including a digital twin partnership with a leading EDA vendor, allowing OEMs to simulate and validate chiplet configurations before tapeout, another tool in the fight to reduce ASIC costs and de-risk development.
Open Source Is Inevitable
We don’t claim open source will dominate every corner of automotive silicon but it will always be there. Just as it has in software, open source will gain permanent, growing market share. The benefits o lower cost, increased transparency, improved security, and multi-sourcing are too powerful to ignore.
The current model is broken. OEMs can no longer afford to shoulder the cost of full ASIC development. Nor can they continue buying black-box silicon that doesn’t align with their platform strategies. The result? Soaring vehicle costs, industry-wide layoffs, and mounting pressure to change course.
Open source chiplets offer a way out through a shared foundation for innovation, flexibility, and sustainability.
At ChipFlow, we’re building that foundation.
