There’s a lot of bad info out there about 5G/6G optics and how companies like Lumentum are building our future network infrastructure. If you want to get a real handle on where connectivity is headed, you have to look past the hype and see what’s actually happening with the technology.
Key Takeaways
- Lumentum’s indium phosphide (InP) is the key material for the 800G-and-up transceivers that next-gen data centers need to function.
- Moving to 6G isn’t just about speed. It’s about building optical parts that can handle terabit-per-second traffic and terahertz frequencies which is something current silicon photonics struggles with.
- Integrated photonics, cramming multiple optical jobs onto one chip, is how we’ll get the efficiency gains and lower latency that future networks demand.
- The market for this stuff is set to explode. LightCounting projects the optical transceiver market by itself will top $20 billion by 2029.
- Open RAN doesn’t kill proprietary gear, it just shifts the demand to standardized, high-performance optical modules for fronthaul and midhaul connections.
Myth 1: Silicon Photonics Will Solve All 5G/6G Optical Challenges
Lots of people think silicon photonics is the silver bullet for all things optical in 5G and 6G because it’s cheap to scale. That’s a huge oversimplification. The idea that a single material can handle every part of these complex networks just isn’t true. Silicon photonics has been great for data center interconnects, getting us to 400 Gbps and even 800 Gbps, but it hits a wall when you need more speed, longer reach, or specific functions for next-gen wireless. The reality is that getting silicon to generate and amplify light efficiently is still a massive engineering problem. That’s where a material like indium phosphide (InP) is still absolutely necessary. InP-based lasers and modulators are simply better at pumping out optical power, tuning wavelengths, and converting signals, all of which are table stakes for 800 Gbps+ transceivers in coherent optical systems. Lumentum has been working with InP for decades, and that expertise is the foundation of its high-performance coherent modules. These modules keep the signal clean over huge distances and across messy network paths, a job that silicon photonics can’t do by itself. A 2024 market analysis from Cignal AI (https://cignal.ai/category/reports/) confirms this, showing that InP parts will continue to dominate high-speed and long-haul applications for the rest of the decade. Silicon photonics won’t disappear, but it’s not going to take over either. The real trend is heterogeneous integration: we’re combining the best of different materials, like InP and silicon, on a single package to get the performance we need for a specific job at the right cost.
Myth 2: 5G/6G Optics are Just Faster Versions of Previous Generations
It’s easy to think that 5G and 6G optics are just 4G tech on steroids, but that view completely misses the massive architectural shifts happening in the network. We’re changing everything about how traffic is managed and how low latency is achieved. This is about building a smarter, more distributed network, and the optics have to change to support it. The demands of 5G, with its mix of mobile broadband (eMBB), low-latency communications (URLLC), and massive IoT (mMTC), forced us to rethink the entire optical layer. For example, we’ve disaggregated the radio access network (RAN) into central units (CUs), distributed units (DUs), and radio units (RUs), creating new fronthaul and midhaul segments. These new links require specific optical transceivers with tight timing synchronization, low power draw, and the ability to survive in an outdoor cabinet. Lumentum’s portfolio has specialized 25G and 100G modules built exactly for these tough 5G fronthaul jobs. Looking ahead, 6G will push things into another dimension, demanding terabit-per-second (Tbps) links and opening up new terahertz (THz) frequency bands. This will require completely new ways of converting signals between the optical and electrical domains. We’re talking advanced modulation formats, photonic circuits that do signal processing on-chip, and maybe even quantum components for security. This isn’t just a fantasy. The International Telecommunication Union (ITU) (https://www.itu.int/en/ITU-T/focusgroups/fg-ai4thz/Pages/default.aspx) is already defining the groundwork for 6G and pointing to these very shifts in optical tech.
Myth 3: Optical Component Manufacturing is a Commodity Business
Anyone who thinks making high-performance optical components is a commodity business has never been inside a fab. It’s a world away from churning out generic electronic parts where price is the only thing that matters. This is a highly specialized field. The fabrication process for advanced optics uses techniques like molecular beam epitaxy (MBE) to grow perfect semiconductor layers atom by atom and electron-beam lithography to etch patterns at the nanoscale. Getting good yields out of these processes is incredibly difficult, and the expertise required is built over decades. Companies like Lumentum pour money into R&D and their manufacturing facilities to stay ahead. For example, building a high-power, tunable laser for a coherent system requires insane control over material purity and packaging just to make sure the thing works reliably for years in the field. This is nothing like a commodity product. On top of that, the IP around these designs creates a huge barrier to entry for newcomers. Consistently shipping high-quality, high-performance parts at scale, while working through a tangled global supply chain, is proof that this is a specialist’s game. Reports from industry analysts like Yole Group (https://www.yolegroup.com/category/photonics-sensing/) back this up, showing the heavy R&D spending and specialized manufacturing that defines this market.
Myth 4: Open RAN Reduces the Need for High-Performance Proprietary Optics
The move to Open RAN (Radio Access Network) is often misinterpreted. People hear “open” and think it means we can use simpler, generic optics. They’re wrong. While Open RAN does standardize the software and hardware interfaces, it actually *increases* the pressure on the optical components. In an Open RAN setup, the fronthaul and midhaul links between the RU, DU, and CU are standardized, often using protocols like eCPRI over Ethernet. This means every optical transceiver connecting those boxes must strictly conform to industry specs for speed, latency, power, and form factor. This creates a bigger market for interoperable optical modules from multiple vendors, but it also means you need serious testing and certification to ensure all the different pieces actually work together when you plug them in. A 25 Gbps SFP28 transceiver for an Open RAN fronthaul, for example, has to meet the exact standards defined by groups like the Optical Internetworking Forum (OIF) (https://www.oiforum.com/) or the IEEE. Leading suppliers like Lumentum are deeply involved in these standards bodies, offering a whole portfolio of compliant transceivers that make Open RAN possible. The complexity just shifts from dealing with one company’s proprietary system to ensuring interoperability across a whole bunch of suppliers, which puts an even greater premium on component quality and standards compliance.
Myth 5: 6G is Too Far Off to Impact Current Optical Development
Thinking 6G is some distant sci-fi concept is a great way to make your current optical R&D obsolete fast. This misconception can lead to some really bad, short-sighted strategic decisions. The truth is, the foundational work for 6G is happening right now, and the optical technologies being developed today for advanced 5G are its direct precursors. The switch between wireless generations is a continuous evolution, not a hard cutover. The core ideas of 6G, integrating artificial intelligence (AI) at the network edge, adding advanced sensing, enabling truly immersive XR, all demand bandwidth and latency that today’s networks can’t deliver. We’re talking about sub-millisecond latency and extreme reliability. This forces the optical layer to evolve towards terabit-scale access networks and ubiquitous optical sensing. So, is anyone actually working on this? Yes. Research into terahertz photonics, integrated optics for AI, and co-packaged optics (CPO) is all happening because of the expected demands of 6G. Companies like Lumentum are deeply involved, knowing that the high-power, widely tunable lasers they’re perfecting today for coherent systems will be the bedrock of flexible, high-capacity transport in a 6G world. If you ignore the 6G roadmap, the optical gear you’re developing today will be a dead end. Flagship 6G initiatives like Europe’s Hexa-X (https://hexax.eu/) project explicitly state how critical advanced photonics are, showing that 6G is already shaping today’s R&D priorities. So, the evolution of 5G and 6G depends entirely on the interplay between what wireless needs and what optics can deliver. Getting these myths straight helps you make better decisions about network investments and technology roadmaps.
How are the optical needs of 6G different from 5G?
5G needs a lot of bandwidth and low latency, but 6G blows those requirements out of the water. It will demand terabit-per-second speeds, work with terahertz frequencies, and embed sensing capabilities everywhere. This requires much more advanced and integrated photonics, and maybe even quantum-safe communication.
Why is indium phosphide (InP) still so important if silicon photonics is getting better?
InP is just better at the physics of generating and amplifying light. For the high-speed, long-distance coherent transceivers running at 800 Gbps and faster, you need the high power and efficiency that InP provides. Silicon photonics is not there yet when it comes to creating light or providing strong optical gain.
How does Open RAN change the market for optical parts?
Open RAN breaks up the network into more pieces, which actually increases the demand for standardized, high-performance optical transceivers to connect everything. Since networks can now mix and match equipment from different vendors, the optical links connecting them have to be top-quality and fully interoperable.
What does ‘heterogeneous integration’ mean for optics?
It’s about taking different materials, like silicon and indium phosphide, and combining them onto a single chip or in one package. The goal is to use the best properties of each material for a specific job, like using silicon for the logic and InP for the laser, to get the best performance and cost.
Will optical components become cheap commodities?
No. The advanced components needed for 5G and 6G are incredibly complex to design and manufacture, requiring huge investments in R&D and specialized fabs. While some very basic, older parts might see price pressure, the high-performance optics driving modern networks will remain a specialized field built on deep expertise.
