OpenGMSL is Analog Devices' plan to turn its proprietary GMSL (Gigabit Multimedia Serial Link) automotive SerDes into a documented, multi-vendor standard, run by a new non-profit, the OpenGMSL Association. It can succeed if other chipmakers actually ship compatible silicon, if the association's compliance testing proves real interoperability, and if governance is seen as independent of ADI. Its advantage is an enormous installed base; its risk is that OEMs see little reason to change a link that already works.
In the automotive industry, few connectivity technologies have achieved the adoption and trust that Maxim's (now Analog Devices') GMSL2 and GMSL3 have earned. ADI reports more than a billion GMSL chips shipped (the original version of this post cited over 1.1 billion links), making GMSL the de facto high-speed backbone for cameras, displays and sensors in many modern vehicles.
Key takeaways
- OpenGMSL formalises the existing GMSL stack (physical, link and protocol layers) so that more than one vendor can build compatible parts.
- It builds on GMSL2 and GMSL3 rather than replacing them, so backward compatibility is the selling point.
- GMSL2 runs at 3 or 6 Gbps forward and GMSL3 at 12 Gbps forward, each with a 187.5 Mbps reverse channel, per ADI product documentation.
- The association was reported to use RAND (reasonable and non-discriminatory) licensing with mandatory compliance testing.
- Competition comes from MIPI A-PHY, the Automotive SerDes Alliance (ASA) Motion Link and TI's FPD-Link.
Now ADI is taking a bold step: opening that very technology. The initiative, called OpenGMSL, seeks to transform GMSL from a proprietary interface into a fully documented, multi-vendor and forward-compatible open ecosystem. The OpenGMSL Association was launched as a non-profit consortium in mid-2025 (press coverage dates the formal launch to June 2025), with ADI as a founding member alongside automotive suppliers, chipmakers and test-equipment companies.
The question is not whether OpenGMSL can replace GMSL2 and GMSL3: it includes them. The real question is whether it can evolve the technology and sustain its position in a fast-changing automotive SerDes market.
From proprietary to open: the rationale
Automotive OEMs prize proven, validated and robust technologies. GMSL3 and GMSL2 earned their place by meeting strict automotive EMC and reliability requirements and by delivering predictable performance across link lengths and environmental conditions.
Yet this success created its own limitation: vendor dependency. With ADI as the sole source of GMSL-compliant silicon, cost, supply and innovation cycles were tied to one roadmap. The 2021 to 2022 chip shortage made single-source parts a board-level concern for many carmakers.
OpenGMSL is ADI's strategic response: a move toward collaboration and openness rather than defending a closed format. It aims to:
- Publish GMSL2 and GMSL3 specifications under an open framework.
- Enable interoperability across multiple silicon vendors.
- Encourage innovation through standardised extensions (higher bandwidth, better EMC, lower power).
- Preserve backward compatibility with existing GMSL2 and GMSL3 systems.
This is less about disruption and more about evolution through openness.
How a GMSL link works
A GMSL link joins a serializer, for example in a camera module, to a deserializer in an ECU or SoC board over a single coax or shielded twisted pair. The forward channel carries high-rate video and data, typically MIPI CSI-2 from an image sensor. A low-rate reverse channel on the same cable carries control traffic (I2C, GPIO) back to the sensor, and power can be delivered over the same coax. This one-cable design is what makes it cheap to route cameras around a car.
A worked bandwidth check, using illustrative figures: an 8-megapixel camera at 30 frames per second and 12 bits per pixel produces about 8.3 million x 30 x 12, or roughly 3.0 Gbps of raw pixel data. With blanking and protocol overhead, that fits a 6 Gbps GMSL2 link with margin, while two such cameras aggregated on one link would need GMSL3 at 12 Gbps.
Technical overview: what changes with OpenGMSL
GMSL3 already supports high data rates and robust EMC performance. OpenGMSL expands the field in several ways:
| Feature | GMSL2 | GMSL3 | OpenGMSL (as proposed) |
|---|---|---|---|
| Forward link rate | 3 or 6 Gbps | 12 Gbps | Includes GMSL2/3 rates; extensions planned |
| Reverse link rate | 187.5 Mbps | 187.5 Mbps | Open, standardised uplink |
| Encoding / modulation | 8b/10b (as originally stated, not verified) | 64B/65B + PAM4 (as originally stated, not verified) | Flexible, standardised profiles |
| EMC / PSD compliance | Proven | Proven | Revalidated under an open test suite |
| Vendor ecosystem | Single (ADI) | Single (ADI) | Multi-vendor (goal) |
| Licensing | Proprietary | Proprietary | Specification under RAND terms, as reported |
Corrections to the original table: the forward rates were listed as about 3 Gbps for GMSL2 and about 6 Gbps for GMSL3; ADI documentation gives 3 or 6 Gbps for GMSL2 and 12 Gbps for GMSL3. Licensing was listed as open and royalty-free; reports of the association's launch describe RAND licensing, which allows reasonable royalties. ADI does not publish line coding details in its public datasheets, so the encoding row is kept as originally written and flagged.
Technically, OpenGMSL does not overhaul the GMSL architecture. It introduces standardised compliance layers, open PHY specifications and a transparent certification process. This opens the door for second-source silicon suppliers, while OEMs keep backward compatibility with existing GMSL links.
The competitive field
| Link | Owner or body | Model |
|---|---|---|
| GMSL2 / GMSL3 | Analog Devices | Proprietary, large installed base |
| OpenGMSL | OpenGMSL Association | Multi-vendor specification built on GMSL |
| FPD-Link III / IV | Texas Instruments | Proprietary |
| MIPI A-PHY | MIPI Alliance (also IEEE 2977) | Open standard |
| ASA Motion Link | Automotive SerDes Alliance | Open standard |
The open standards had a head start on governance but lack GMSL's installed base. OpenGMSL tries to combine the two: an installed base with an open specification.
Challenges ahead
1. Ecosystem inertia
GMSL3 and GMSL2 are deeply embedded across camera modules, ECUs and SoCs. OEMs may hesitate to re-qualify links or invest in new validation workflows unless clear cost or sourcing benefits are shown.
2. Multi-vendor coordination
An open standard must deliver real interoperability, which means shared compliance tests, certification and version control. If second-source vendors interpret the specification differently, interoperability risks could outweigh the benefits of openness.
3. Technical parity
OpenGMSL must maintain, or exceed, the performance envelope of GMSL3. Keeping low jitter, deterministic latency and EMC margins across independent implementations will need rigorous coordination and testing.
4. Governance and trust
For OEMs to embrace OpenGMSL, governance must be genuinely collaborative, not a de facto ADI-controlled process. The association was reported to have its own independent board; how decisions are made in practice will matter more than the structure on paper.
Opportunities: building a sustainable ecosystem
- Reduced supply chain risk: multi-vendor sourcing reduces the risk of shortages or price swings.
- Lower cost of innovation: standardised interfaces speed up integration of sensors and displays from different suppliers.
- Interoperability: simplifies design reuse across platforms and generations.
- Future-proofing: open extension frameworks allow scaling to higher speeds and new protocols without breaking legacy links.
OpenGMSL could thus become the "Linux of automotive SerDes": a stable, extensible, community-maintained backbone that grows with the industry.
What adopters should evaluate
- Compliance and certification: how mature are the open compliance test suites, and who governs updates?
- Performance consistency: do early multi-vendor prototypes match the EMC and bit error rate (BER) performance of legacy GMSL3 links?
- Backward compatibility: can existing GMSL-based modules integrate seamlessly with OpenGMSL parts?
- Cost and licensing: are there real BOM or NRE savings from open participation once RAND royalties are counted?
- Longevity: what is the roadmap beyond 12 Gbps, and who commits to long-term support?
Strategies for OpenGMSL to succeed
- Establish transparent governance: run the association like MIPI or ASA, with open processes for revisions, compliance and IP policy.
- Preserve backward compatibility: OEMs should move from GMSL3 to OpenGMSL-certified designs without hardware redesign.
- Enable multi-vendor certification: define compliance suites early and certify several PHY vendors to avoid any perception of single-vendor control.
- Demonstrate performance: benchmark open implementations under real-world EMC and latency tests to prove parity with proprietary GMSL3.
- Build an ecosystem story: publish reference designs, open evaluation boards and automotive-grade test reports to attract Tier 1 interest.
FAQ
Is OpenGMSL compatible with existing GMSL2 and GMSL3 parts?
That is its stated aim: the specification is built on GMSL2 and GMSL3, so certified parts are meant to interoperate with the installed base.
Is OpenGMSL royalty-free?
Not as reported at launch. The association described RAND (reasonable and non-discriminatory) licensing terms, which may include royalties.
How does it differ from MIPI A-PHY?
A-PHY was designed as an open standard from the start, with its own PHY. OpenGMSL opens up an existing, widely deployed proprietary link. The choice for OEMs is between a clean-sheet standard and continuity with what they already ship.
Outlook: evolving, not competing
OpenGMSL's biggest advantage is also its biggest challenge: it carries the legacy of its own success. By extending rather than replacing GMSL2 and GMSL3, it builds on proven reliability while promising openness, flexibility and vendor diversity.
If ADI can show genuine multi-vendor engagement, transparent governance and equal or better performance by 2027, OpenGMSL could move from a proprietary SerDes link to an industry foundation for open automotive connectivity.
Otherwise, it risks being remembered as an ambitious self-reflection, a legacy made visible, rather than a revolution.
Author: Murugavel, exploring semiconductor innovation, open standards and the future of automotive electronics.


There is a lot of speculation on this topic. You have captured some of the key concerns very well.
ReplyDeleteIn my opinion, OpenGMSL is actually adding more confusion in the market and not sure what real cost and performance advantages it's offering to replace incumbent GMSL & FPD-Link solution. Without that clear benefits, adoption by multiple vendor and OEM is going to be tough nut to crake.
ReplyDelete