Optical photonics is no longer confined to long-distance fiber networks. It is becoming increasingly relevant wherever high-speed data movement, energy efficiency and precision matter: data centers, AI infrastructure, telecommunications, sensing and emerging wireless systems.
Recent industry discussion around Nokia’s interest in data-center optical networking illustrates why. The company reportedly sees an opportunity to extend beyond coherent optical transport and data-center interconnects toward optical solutions used inside data centers. Its existing laser capabilities are an important part of that conversation.
Why AI changes the network
Artificial intelligence increases the volume of data moving between processors, memory, storage and networking equipment. As AI systems scale, the connections between compute resources can become as important as the processors themselves.
Traditional pluggable optical transceivers have served data-center networks well, but they can consume meaningful power. Industry analysis cited in recent reporting estimates that digital signal processors may account for roughly 30% to 40% of pluggable-transceiver power consumption.
That has increased interest in approaches that reduce electrical conversion and DSP dependence. Near-packaged optics places optical engines closer to the processor on the same board. Co-packaged optics goes further by integrating optical engines with the processor substrate.
Scale-out and scale-up
Most current discussion of near-packaged and co-packaged optics concerns scale-out networking: connecting servers, racks and clusters across a data center.
Scale-up networking is a different challenge. It concerns high-bandwidth connections within or closely around accelerated computing systems, where processor density, memory access and latency can become critical constraints.
Optical approaches may eventually support both. However, the reported market view is that scale-out applications are currently clearer than scale-up applications. That distinction matters for buyers evaluating claims about optical interconnects.
Lasers remain foundational
Lasers are essential components in many optical systems. They generate the light that carries information through fiber, optical interconnects and sensing systems.
A company with established laser capability has an advantage when entering adjacent optical markets. But lasers alone do not determine success. Packaging, alignment, thermal management, manufacturing yield, reliability and system integration remain essential.
The reported Nokia discussion reflects that reality. The company is reportedly evaluating whether to supply external lasers, build complete near-packaged-optics solutions, pursue co-packaged optics or some combination of those approaches.
NPO and CPO are not interchangeable
Near-packaged optics and co-packaged optics address related problems, but they involve different engineering trade-offs.
Near-packaged optics may offer advantages in serviceability because certain components can potentially be replaced if they fail. Co-packaged optics may reduce some distances between optical and electronic functions, but it can create more demanding integration and packaging requirements.
There is no universal answer yet. The right approach depends on the workload, system architecture, reliability requirements, manufacturing maturity and total cost of ownership.
Beyond data centers
The growth opportunity for photonics extends beyond AI data centers.
Optical technologies support telecommunications transport, fiber-to-the-home networks, sensing systems, automotive lidar, industrial inspection, medical diagnostics and scientific instruments. Integrated photonics can combine multiple optical functions on a chip, potentially reducing size and simplifying some assembly tasks.
But integration does not automatically produce a better product. Packaging, testing, yield and system-level performance still determine whether a photonic solution delivers value.
What decision-makers should ask
For organizations evaluating optical photonics, the useful questions are practical:
What problem does the technology solve? Where does it fit in the system? What evidence supports its performance claims? How mature is manufacturing? What are the reliability, serviceability and integration requirements?
Those questions matter more than a general claim that photonics is the future.
A measured opportunity
Photonics is an important enabling technology for high-speed communications, AI infrastructure and advanced sensing. Its applications are expanding, and interest from established telecommunications suppliers illustrates the market’s potential.
At the same time, commercial adoption depends on more than enthusiasm. The companies most likely to benefit will be those that can connect optical capability to a clear system-level need, reliable manufacturing and a credible path to deployment.
Editorial basis: Reporting by Fierce Network on Nokia’s stated interest in data-center optical networking, near-packaged optics, co-packaged optics and laser supply. This article is educational and does not endorse any company, product or investment.