The fiber optic component industry has operated on a remarkably stable model for decades: fiber manufacturers make fiber, connector companies make connectors, transceiver makers assemble transceivers, and system vendors buy these discrete components to build their equipment. Co-packaged optics (CPO) threatens to collapse several of these layers into one, with profound implications for every company in the supply chain.
What CPO Actually Means for Component Makers
Co-packaged optics integrates photonic engines directly onto the same package substrate as the switch ASIC or network processor. Instead of plugging discrete transceivers into a front-panel faceplate, optical engines sit millimeters from the electronic chip, connected by short electrical traces rather than power-hungry SerDes links running across a PCB.
For component manufacturers, CPO changes almost everything about how optical components are sourced, assembled, and delivered. The pluggable transceiver — that self-contained module with its own housing, electrical interface, optical engine, and fiber connector — may not disappear entirely, but CPO creates an alternative pathway where many of those components are consumed in fundamentally different ways.
The Winners and Losers
Photonic chip makers stand to gain. Companies like Broadcom, Intel, and Ayar Labs that can deliver bare photonic dies or chiplets for integration onto multi-chip modules are well-positioned. The volume per socket increases (a single switch might consume 32-64 optical engines), and the integration becomes tighter with the ASIC vendor's packaging ecosystem.
Traditional transceiver ODMs face disruption. Companies like InnoLight, Hisense Broadband, and Source Photonics have built their businesses on assembling pluggable modules — procuring laser chips, driver ICs, TIAs, lenses, and housings, then integrating them into tested modules. In a CPO world, the ASIC vendor or system OEM may vertically integrate optical engine assembly, cutting out the transceiver ODM.
Connector companies face a mixed outlook. On one hand, CPO eliminates some pluggable connector interfaces. On the other hand, the fiber management within CPO-equipped systems still requires connectors — and the precision demands may actually increase as fiber routing becomes more constrained. Companies like US Conec and Senko are developing new connector form factors specifically for CPO applications.
Precision component suppliers may actually benefit. The laser chips, photodetectors, micro-lenses, and waveguides that go into a CPO photonic engine are the same fundamental components used in pluggable transceivers — often with tighter specifications. Suppliers of these precision components (Lumentum, Coherent, and specialty micro-optics firms) may see demand increase even as the packaging architecture changes.
The Assembly Challenge
Perhaps the most significant impact of CPO on the component supply chain is the dramatic increase in assembly complexity. A pluggable transceiver is assembled and tested as a standalone unit. If it fails testing, it is scrapped at relatively low cost. A co-packaged optical engine, however, is integrated onto a multi-chip module alongside a multi-billion-transistor ASIC — a package that might cost thousands of dollars.
This changes the calculus of component quality. Known-good die (KGD) testing becomes essential for photonic engines before they are integrated. Optical testing at the die level — verifying laser performance, modulator response, and photodetector sensitivity before packaging — requires new test equipment and methodologies that the industry is still developing.
The assembly process itself requires capabilities that few companies possess today. Photonic engine attachment demands placement accuracy of ±1 µm or better, thermal management of optical components with different requirements than digital ASICs, and fiber coupling that must survive the thermal cycling of the entire package. This is not a task for standard SMT assembly lines.
Supply Chain Consolidation Is Inevitable
CPO will likely accelerate the vertical integration trend that is already reshaping the optical networking industry. When the ASIC vendor controls the photonic engine specification and the system OEM controls the package design, the traditional model of sourcing components from a fragmented supply base becomes untenable.
Expect to see ASIC vendors (Broadcom, Marvell, potentially NVIDIA) acquiring or deeply partnering with photonic integration companies. System OEMs (Microsoft, Google, Meta, Amazon) will continue building internal optical engineering teams to co-design CPO solutions with their ASIC partners. The transceiver ODM, historically the integrator of optical components, may need to pivot toward becoming a photonic engine assembler — or risk being disintermediated.
The Timeline Is Longer Than You Think
Despite the hype, mass deployment of CPO remains years away. The industry consensus puts meaningful CPO volume in the 2027-2029 timeframe, with pluggable transceivers continuing to dominate through at least 2030. The reasons are practical: pluggable transceivers work, they are field-replaceable, they come from a mature supply chain, and the economic case for CPO only becomes compelling at 1.6T per port and beyond.
But component manufacturers cannot afford to wait. The development cycles for new packaging technologies, precision assembly equipment, and tested-good-die methodologies are measured in years, not months. Companies that are not investing in CPO-capable manufacturing today will find themselves locked out of the supply chain when volume deployment arrives.
What This Means for the Industry
The fiber optic component supply chain is entering a period of structural change not seen since the transition from multimode to single-mode in the metro network. CPO will not eliminate the need for precision optical components — if anything, it will increase demand for higher-performance variants. But it will fundamentally reshape who assembles those components, how they are tested, and who captures the value in the supply chain.
For component manufacturers, the strategic imperative is clear: invest in precision, invest in testability, and invest in partnerships with the ASIC and system companies that will define the CPO ecosystem. The companies that thrive will be those that can deliver known-good optical components at the quality levels that multi-thousand-dollar packages demand.