Ethernet transceiver market to hit $22B by 2035 on AI demand
Analyst firm IndexBox forecasts the global Ethernet transceiver market will grow from $12 billion in 2025 to over $22 billion by 2035, driven by hyperscale data center upgrades to 400G and 800G links. This infrastructure shift supports the intensifying bandwidth and low-latency requirements for modern high-resolution video streaming and AI/ML workloads.
Key Takeaways
- High-speed segments (400G+) are projected to account for 40% of total revenue by 2030, rising from 15% in 2025.
- Asia-Pacific currently produces 75% of global unit output, while North America and Europe represent 55% of consumption.
- Industrial and automotive Ethernet subsegments are expected to grow 8-10% annually through 2035.
- The 800G interface is anticipated to contribute 25-30% of data center transceiver revenue by 2030.
Why It Matters
The transition to 800G and 1.6T interfaces is no longer a luxury but a requirement for the data-heavy AI and high-resolution video streaming workloads defining the next decade. For the streaming ecosystem, this indicates a structural shift where networking costs may begin to rival compute expenses as hyperscalers exhaust existing 100G capacity. Short-term supply chain concentration for critical photonic components like laser chips and DSPs remains a high-risk factor that could delay volume adoption by up to 18 months if new fabrication capacity in the US and Taiwan falls behind. Watch for the emergence of co-packaged optics (CPO) and silicon photonics, which are projected to capture 10-15% of new deployments by 2030 as operators seek power-saving solutions.
Additional Context
The Ethernet market is undergoing a fundamental realignment as it displaces proprietary interconnects in high-performance computing. Per Dell'Oro Group (July 2025), Ethernet is effectively winning the 'war' against InfiniBand in AI back-end networks, a sector that was 80% dominated by InfiniBand as recently as late 2023. This rapid shift is driven by the maturation of the Ultra Ethernet Consortium (UEC 1.0) specifications in June 2025, which allow Ethernet to approach InfiniBand’s sub-2 microsecond latency while leveraging a more open ecosystem. Dell'Oro projects that Ethernet could drive nearly $80 billion in data center switch sales over the next five years as hyperscalers seek to reduce vendor lock-in. Nvidia has emerged as a surprising dominant force in this shift. According to IDC (June 2026), Nvidia’s data center Ethernet switch revenue grew 192.7% year-over-year to $2.1 billion, giving it a 21.5% market share—surpassing longtime leaders like Arista Networks (20.7%). This growth is largely credited to the Spectrum-X platform, which hyperscalers like Meta and Oracle use to interconnect millions of GPUs into giga-scale AI factories. The integration of networking directly with GPU clusters is altering traditional procurement, where transceivers and switches are now purchased as a unified platform rather than separate infrastructure layers. Technological roadmaps are also accelerating to keep pace with next-generation silicon. Per Cignal AI (April 2026), 800G transceiver shipments achieved a 100% year-on-year increase in 2025, reaching institutional volume. Attention is now shifting to 1.6T modules, which entered advanced production for Nvidia and other hyperscale applications in late 2025. While pluggable optics remain the standard for now, industry analysts expect CPO and linear pluggable optics (LPO) to gain significant traction by 2028-2030 to mitigate the massive power consumption required to drive Terabit-scale bandwidth per rack.
Read full article at indexbox.io
Get this in your inbox → Subscribe
Enjoy our coverage?
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source