Macroplane › Blog › Co-Packaged Optics (CPO): The Photonic Switches Rewiring AI Networks (2026)

Co-Packaged Optics (CPO): The Photonic Switches Rewiring AI Networks (2026)

2026-06-10

A deep dive into co-packaged optics (CPO) — NVIDIA Spectrum-X Photonics switches (SN6810 102.4 Tbps, SN6800 409.6 Tbps, Spectrum-6 ASICs), the ~3.5x power and ~10x reliability gains over pluggable transceivers, why they ship liquid-cooled, the 2025–2026 CPO news timeline, and the full supply chain making them: $NVDA, $AVGO, $TSM, $COHR, $LITE, $FN, $POET, plus the LPO/AEC camp ($CRDO, $ALAB).

More Macroplane supply-chain guides · Baskets · Investment theses · Macro trends

Co-Packaged Optics (CPO): The Photonic Switches Rewiring AI Networks (2026)

Co-packaged optics (CPO) move the optical engine out of a plug-in module and onto the same package as the switch chip — and that one change is the biggest shift in data-center networking in a decade. As AI clusters scale past hundreds of thousands of GPUs, the network that connects them is hitting a power and reliability wall that pluggable transceivers can't climb. CPO is the answer the whole industry is now racing to ship. This is a deep dive into what these photonic switches — NVIDIA's Spectrum-X and Quantum-X Photonics, Broadcom's CPO line — actually do, how much better they are, why they arrive liquid-cooled, and — for investors — exactly who is making them.

👉 See it in context: the CPO supply chain lives in the Photonics / CPO basket on Macroplane — $COHR, $LITE, $FN, $POET, $AAOI and the optical names — and the broader Silicon Photonics & Optical Interconnects trend.

🎥 Watch: NVIDIA's New Photonic Technology Explained — a clear visual walkthrough of how co-packaged optics works.

What a co-packaged optics switch actually is

In a conventional AI switch, the chip in the middle is an electrical ASIC, and light only happens at the edges: each port has a pluggable optical transceiver (a QSFP/OSFP module you can pull out by hand) that converts the switch's electrical signals into laser light for the fiber. Every one of those modules carries its own laser, its own digital signal processor (DSP), and its own retimer — and at 800G and 1.6T, those modules are getting hot, power-hungry, and failure-prone.

Co-packaged optics deletes the module. The optical engines are integrated directly onto the switch ASIC's package, so light is generated millimeters from the silicon instead of inches away across a circuit board and a connector. NVIDIA's Spectrum-X Photonics line is the clearest example of the new design:

  • NVIDIA SN6810 — 102.4 Tbps of throughput (512×200GbE, 256×400GbE, or 128×800GbE), a single Spectrum-6 ASIC, MMC fiber connectors on the faceplate, ~9 W of optics per 1.6T port (versus 25–30 W for a pluggable), in a 2RU liquid-cooled chassis.
  • NVIDIA SN6800 — 409.6 Tbps (512×800GbE), four Spectrum-6 ASICs in one liquid-cooled chassis.

Both replace the pluggable-transceiver front panel with silicon-photonics co-packaged optics for low-latency, single-mode fiber connectivity. On the InfiniBand side, NVIDIA's Quantum-X Photonics Q3450-LD does the same thing for the back-end GPU fabric: 115.2 Tbps across 144 ports of 800G, four Quantum-X800 ASICs, 18 field-replaceable external laser modules, and a measured 3.95 kW — versus roughly 7 kW for the equivalent pluggable-based switch. The common thread: the optics and the switch silicon now share one package.

The timing matters for investors: all of this was announced at GTC in March 2025, Quantum-X Photonics started shipping in early 2026 at low volume, and NVIDIA began shipping the first Spectrum-X CPO switches to select partners in June 2026, with the production ramp scheduled for the second half of the year.

The benefits of co-packaged optics — how much better is it?

The gains aren't incremental. By co-packaging the optics, NVIDIA cites four step-changes versus traditional pluggable-transceiver networking:

  • ~3.5× better power efficiency in the optics — the single biggest lever at cluster scale.
  • ~63× better signal integrity — the electrical path from ASIC to optics drops from ~22 dB of loss to ~4 dB.
  • ~10× better network resiliency — 4× fewer lasers, higher-quality and externally mounted instead of thousands of in-module ones.
  • ~1.3× faster time to deployment — fewer discrete parts to assemble, test, and debug.

Two of those deserve a closer look:

Power. A large AI fabric can burn through megawatts just on its optical transceivers. A 1.6T-generation pluggable draws ~25–30 W (the figure behind Jensen Huang's GTC keynote math; today's 800G modules run ~13–17 W), and a 400,000-GPU cluster needs hundreds of thousands of optical connections. NVIDIA's framing is blunt: at hyperscale, CPO saves tens of megawatts of networking power — power that can instead be spent on GPUs. When the binding constraint on AI is electricity (see our data-center power and nuclear coverage), shaving the optics power bill is a direct lever on how many accelerators fit in a given site.

Reliability. The laser is the single most common failure point in optical networking. CPO designs use far fewer lasers, run them as high-power **external light sources** that can be serviced without touching the switch, and keep them off the hot ASIC package. Fewer lasers, better cooled, in a more controlled environment is what gets you to that ~10× resiliency claim — and at the scale of a single training run spanning 100,000+ GPUs, a network that stays up is worth as much as one that's fast.

CPO switchThroughputPortsSwitch ASICsPowerForm factor
NVIDIA SN6810 (Spectrum-X)102.4 Tbps128×800GbE1× Spectrum-6~9 W optics per 1.6T port2RU, liquid-cooled
NVIDIA SN6800 (Spectrum-X)409.6 Tbps512×800GbE4× Spectrum-6~3.5× better than pluggableliquid-cooled
NVIDIA Quantum-X Photonics (Q3450-LD)115.2 Tbps144×800G InfiniBand4× Quantum-X8003.95 kW (vs ~7 kW pluggable-based)4U, liquid-cooled
Broadcom Bailly (Tomahawk 5 + CPO)51.2 Tbps128×400G FR41× Tomahawk 5~70% lower optics power4RU, air-cooled (Micas)
Broadcom Tomahawk 6 "Davisson"102.4 Tbps200G/lane optical1× Tomahawk 6~70% lower optics powershipping since Oct 2025

*Figures as published by the vendors — each switch name links to its primary source (NVIDIA's engineering blog, Lambda's Q3450-LD teardown, Broadcom's announcements). Treat them as marketing-grade specs, not audited numbers.*

Copper vs photonics: the trend that forces the switch

For two decades copper kept up by brute force — every time signaling doubled (NRZ → PAM4, 25G → 50G → 100G → 200G per lane), the channel got a little smarter. But physics is winning. At 200G-per-lane PAM4, a passive copper cable carries a usable signal less than a meter before the loss is unrecoverable. Optics doesn't hit that wall — a single-mode fiber runs hundreds of meters at the same rate. The catch is power: a pluggable optical module spends a big slice of its energy in the DSP that cleans the signal up across the board and the connector. CPO attacks exactly that — by putting the optics on the package, it can drop the heavy DSP and cut the energy per bit by roughly 3× while keeping optical reach.

Here's the same story as a side-by-side, from in-rack copper to co-packaged optics:

Link typeEnergy/bit*ReachLatencyField-serviceableWhere it fits
Passive copper (DAC)~1–2 pJ/bit<1–2 mlowestyesin-rack, GPU-to-GPU short hops
Active copper (AEC)~5–8 pJ/bit~2–7 mvery lowyescross-rack, NVLink-class scale-up
Pluggable optics (DSP)~15–20 pJ/bit100 m–2 km+lowyestoday's switch front panels
LPO (linear pluggable)~8–10 pJ/bit~100 mloweryesnear-term power cut, DSP removed
Co-packaged optics (CPO)~5–7 pJ/bit100 m–2 km+lowest opticallaser-serviceablehighest-radix AI switches

\Energy figures are rounded, widely-reported industry estimates for direction and magnitude — not audited vendor numbers.*

Why CPO arrives liquid-cooled

Look at the power column again: NVIDIA's Quantum-X Photonics packs 3.95 kW into a 4U box, and the Spectrum-X line pushes the same density further. That is an enormous thermal load for a switch — and the harder problem is that co-packaging deliberately puts the heat-sensitive optics right next to the hottest silicon in the box. Air struggles to move that much heat out of that small a volume while keeping the lasers in their happy temperature range (laser wavelength and lifetime both drift with heat). Broadcom's 51.2T Bailly generation still manages with high-efficiency air cooling in 4RU; NVIDIA's denser 102.4T-per-ASIC designs do not.

So at the leading edge, CPO and liquid cooling ship together. The same direct-to-chip and cold-plate liquid loops that the AI rack already needs for its GPUs get extended to the switch. That makes the cooling supply chain part of the CPO story: the precision-cooling and liquid-distribution names — Vertiv ($VRT) and the rest of the Cooling basket — are downstream beneficiaries of every CPO switch deployed, because each one is another high-density, liquid-cooled box in the row. CPO doesn't just change the optics; it pulls more of the rack onto liquid.

Who's actually making co-packaged optics

This is the part that matters for investors — CPO is a system, and the value is spread across several layers. Here's the map, grounded in the relationships visible on Macroplane.

The switch makers putting optics on the package

The companies that actually build the AI switches — and are racing to co-package the optics onto them:

  • NVIDIA ($NVDA) — Spectrum-X Photonics (Ethernet, Spectrum-6) and Quantum-X Photonics (InfiniBand); the most aggressive CPO roadmap in the market.
  • Broadcom ($AVGO) — the merchant-silicon giant (Tomahawk / Jericho) and NVIDIA's main rival for the CPO socket. Its Bailly CPO switch (Tomahawk 5) has been running in production at Meta ($META) — Broadcom reported a million flap-free CPO port device-hours there in October 2025 — and the Tomahawk 6 "Davisson" follow-on (102.4 Tbps, 200G/lane optics) has been shipping since October 2025, with a 400G/lane generation in development.
  • Cisco ($CSCO) — Silicon One switch silicon plus a decade of in-house silicon-photonics and CPO research (it acquired Acacia for exactly this).
  • Arista Networks ($ANET) — the hyperscaler Ethernet-switch leader; ships merchant silicon with pluggables today and is a gating customer for whichever CPO approach wins.
  • Marvell ($MRVL) — Teralynx switch silicon and optical DSPs, plus Celestial AI's Photonic Fabric (acquired February 2026) and a custom co-packaged / attached-optics roadmap for hyperscaler ASICs; it also sells the DSPs into transceiver makers like $COHR, and in March 2026 NVIDIA took a $2B stake alongside an NVLink Fusion and silicon-photonics partnership.
  • Nokia ($NOK) — IP routing and, via the Infinera acquisition, optical systems and indium-phosphide PICs; more a router and optical-transport play than data-center Ethernet, but in the same photonics fight.

The venture-backed challengers attacking the optical interconnect from a different angle are consolidating fast: Celestial AI (Photonic Fabric) is now inside $MRVL, while Lightmatter (Passage interposer) and Ayar Labs (optical I/O chiplets, fresh off a $500M Series E in March 2026) remain private — and remain who the incumbents are racing.

The silicon-photonics foundry

CPO is built on a photonic integrated circuit, and someone has to fab it:

  • TSMC ($TSM) — its COUPE (Compact Universal Photonic Engine) silicon-photonics platform is the foundry backbone behind NVIDIA's photonics; the same packaging muscle (advanced packaging is the sister trend) that makes CoWoS makes CPO.
  • GlobalFoundries ($GFS) and Tower Semiconductor ($TSEM) — the other merchant silicon-photonics foundries.

The optical engines, modulators & assembly

The light-making and light-shaping layer — and the part of the chain that's most levered to CPO volume:

  • Coherent ($COHR) — lasers, optical components and transceivers; already a primary 800G/1.6T optics supplier to $NVDA, and a laser-die supplier into external-light-source modules. In March 2026 NVIDIA invested $2B in Coherent with multi-billion purchase commitments for laser capacity.
  • Lumentum ($LITE) — lasers and photonics, a core external-laser and component supplier for CPO; it took the matching $2B NVIDIA investment in March 2026.
  • Fabrinet ($FN) — the contract manufacturer that assembles a huge share of the industry's advanced optics (including for $COHR); CPO assembly is exacting, precision work that flows here.
  • POET Technologies ($POET) — its optical interposer is a CPO-native packaging approach for integrating lasers and waveguides.
  • Applied Optoelectronics ($AAOI), MACOM ($MTSI), Lightwave Logic ($LWLG) — lasers, components, and next-gen (polymer) modulator materials.
  • Sivers Semiconductors ($SIVE.ST) — InP distributed-feedback (DFB) lasers and external light sources; the flagship announced DFB-array supplier into Ayar Labs' SuperNova (more in our Sivers deep dive).
  • Ayar Labs (private) — optical I/O chiplets and its SuperNova external light source, with $SIVE.ST the lead announced laser supplier and $LITE, $COHR, and Sumitomo as additional die sources.

The lasers, substrates & fiber

  • External light sources and the III-V epi underneath them — $LITE and $COHR for lasers; IQE ($IQEPY) and AXT ($AXTI) for the InP/GaAs epiwafers and substrates that lasers are grown on (see the InP & Substrates basket).
  • Corning ($GLW) and TE Connectivity ($TEL) — the single-mode fiber and high-density MMC/MPO connectors that CPO cabling depends on.

The "not yet" camp — linear pluggable optics (LPO) and AEC

Not everyone thinks CPO wins immediately. A cohort is betting that linear-drive pluggable optics (LPO) and active electrical cables (AEC) bridge the gap — keeping optics pluggable but stripping out the power-hungry DSP, or replacing short optical links with smarter copper:

  • Credo ($CRDO) — the AEC leader, plus LPO/retimer products.
  • Astera Labs ($ALAB) — PCIe/CXL connectivity and the scale-up fabric; the Networking / Retimers basket is where this camp lives.
  • MaxLinear ($MXL) and others on the DSP/PHY side.

The honest read: CPO and LPO/AEC coexist for now — CPO for the highest-radix switches where power and density are binding, pluggables/LPO/AEC for the rest. The investable point is that all of these roads run through the same optics and connectivity suppliers.

Co-packaged optics news: the timeline that turned CPO real

If you want the one-paragraph version of why 2026 is the year CPO stopped being a roadmap slide, it's this sequence:

  • October 2025 — Broadcom ships Tomahawk 6 "Davisson", the first 102.4 Tbps Ethernet switch with co-packaged optics (Broadcom), and reports a million flap-free CPO port device-hours from Bailly running in production at $META (Broadcom). CPO's reliability argument moves from claim to measured data.
  • February 2026 — Marvell ($MRVL) closes its acquisition of Celestial AI, folding the Photonic Fabric scale-up interconnect into a public company (Marvell 8-K).
  • March 2026 — NVIDIA invests $2B each in Coherent ($COHR) and Lumentum ($LITE) with multi-billion purchase commitments to lock up laser capacity (CNBC), takes a $2B stake in Marvell alongside an NVLink Fusion + silicon-photonics partnership (Marvell 8-K), and Ayar Labs raises a $500M Series E (The Next Platform). The supply chain stops being merchant and starts being spoken for.
  • May 2026 — TSMC makes COUPE the centerpiece of its Technology Forum and pushes the platform toward mass production (Bits&Chips).
  • June 2026 — NVIDIA starts shipping the first Spectrum-X CPO switches to select partners, built on COUPE, with the volume ramp set for the second half of the year (Taipei Times) — and the first public teardown of a Quantum-X CPO unit lands the same week (Lambda).
  • August 2026 — a 19-company coalition led by Lightmatter turns Open Silicon Photonics for AI Systems into an official Open Compute Project workstream, publishing a roughly 300-page architectural blueprint and targeting first specifications in Q4 2026 (DCD, Lightmatter). CPO stops being three vertically integrated programmes and starts becoming a merchant market.

For the live feed of what's moving the names, the Silicon Photonics & Optical Interconnects trend page tracks co-packaged optics news against the companies it touches, daily.

The standard: what Q4 2026 changes

Until August 2026 co-packaged optics was three vertically integrated stacks. NVIDIA built its own optical engines for its own switches. Broadcom built its own on TSMC COUPE. If you were a hyperscaler who wanted CPO, you bought one vendor's entire architecture.

The Open Silicon Photonics for AI Systems workstream is the attempt to break that. Formally launched inside the Open Compute Project on 13 August 2026, it pairs a roughly 300-page architectural blueprint with a target of first specifications in Q4 2026. The engineering goal is to scale OCP Modular Hardware System and Open Rack v3 environments from 72 nodes to more than 1,024 while keeping multi-vendor interoperability.

The membership list is the interesting part, because it tells you who expects to sell into a merchant CPO market: Celestica ($CLS), Dell ($DELL), Flex ($FLEX), Foxconn Interconnect Technology ($6088.HK), Global Unichip ($3443.TW), Hyve Solutions, Keysight ($KEYS), Lightmatter (private), Qualcomm ($QCOM) and Quanta Cloud Technology ($2382.TW), across ten founding and nine new members.

Two consequences for investors:

Component suppliers get addressable. In a single-vendor architecture, the optical engine supplier is whoever NVIDIA or Broadcom chose. In a specified one, laser, amplifier and interposer suppliers can sell to anyone building to the spec. That is what makes the smaller names in layer two of the photonics stack investable rather than hopeful.

Test moves upstream. You cannot field-replace an optical engine sealed inside a switch package, so the yield and validation burden shifts into manufacturing. Keysight's presence as a founding member is the tell, and Celestica's hyperscaler CPO switch award, ramping in 2027, is the manufacturing side of the same trade.

The date to watch is Q4 2026, when the first specifications are due.

The investable map: photonics stocks across the CPO stack

CPO is a classic picks-and-shovels theme: the switch vendors get the headlines, but the durable exposure is in the components every design needs regardless of who wins.

  • The thesis: POET ($POET): The Photonic-Integrated CPO Engine Outlier — the small-cap pure-play built around the co-packaged-optics light engine, with $POET pinned and the live price tracked.
  • The basket: Photonics / CPO — $COHR, $LITE, $FN, $CIEN, $POET, $AAOI, $LWLG and the rest, with live performance.
  • The trends: Silicon Photonics & Optical Interconnects and Advanced Semiconductor Packaging — CPO sits at the intersection of both.
  • The adjacent layers: Networking / Retimers for the LPO/AEC camp, InP & Substrates for the laser materials, and Cooling for the liquid loops CPO forces.
  • The big picture: where optics sits in the whole stack — The AI Supply Chain Map.

What is co-packaged optics (CPO)?

Co-packaged optics integrates the optical engine — the part that converts electrical signals to laser light — directly onto the same package as a switch or accelerator chip, instead of housing it in a separate pluggable transceiver module. This shortens the electrical path, cuts power, and improves reliability, which is why it's being adopted for the highest-bandwidth AI networking switches.

Why does AI networking need CPO?

At 800G and 1.6T per port, pluggable transceivers consume too much power and fail too often to scale to clusters of hundreds of thousands of GPUs. The optical transceivers in a large AI fabric can burn megawatts on their own. CPO addresses both: NVIDIA cites roughly 3.5× better optical power efficiency and ~10× better network resiliency versus pluggables — power and uptime that translate directly into more usable GPUs.

What are the benefits of co-packaged optics?

Four step-changes versus pluggable transceivers, per NVIDIA: ~3.5× better optical power efficiency, ~63× better signal integrity (the ASIC-to-optics electrical loss drops from ~22 dB to ~4 dB), ~10× better network resiliency from 4× fewer, externally mounted lasers, and ~1.3× faster deployment. Broadcom's production data at Meta backs the power claim at roughly 65–70% optics power savings.

When do NVIDIA's CPO switches ship?

Quantum-X Photonics (InfiniBand) began shipping in early 2026 at low volume after slipping from late 2025. The first Spectrum-X Photonics Ethernet switches (SN6810, SN6800) started reaching select partners in June 2026, with the production ramp in the second half of 2026. Broadcom's CPO line has been shipping longer — Bailly since 2024–25 and Tomahawk 6 "Davisson" since October 2025.

Which companies make co-packaged optics?

The switch/integrator layer is led by NVIDIA ($NVDA), with Broadcom ($AVGO), Marvell ($MRVL), and Cisco ($CSCO) competing. The photonics are fabbed by foundries led by TSMC ($TSM), with GlobalFoundries ($GFS) and Tower ($TSEM). The optical engines, lasers, modulators, and assembly come from Coherent ($COHR), Lumentum ($LITE), Fabrinet ($FN), POET ($POET), Applied Optoelectronics ($AAOI), MACOM ($MTSI), and Lightwave Logic ($LWLG), plus private players like Ayar Labs. Fiber and connectors come from Corning ($GLW) and TE Connectivity ($TEL).

Why are CPO switches liquid-cooled?

Because they pack extreme power density — NVIDIA's Quantum-X Photonics is 3.95 kW in a 4U chassis — and co-packaging puts heat-sensitive optics next to the hottest silicon. Air cooling struggles to move that much heat while keeping the lasers in their stable temperature range, so NVIDIA's CPO switches ship with direct liquid cooling, extending the same loops the GPUs already use (Broadcom's lower-density 51.2T Bailly systems remain air-cooled). That makes cooling names like Vertiv ($VRT) downstream beneficiaries.

CPO vs LPO vs AEC — what's the difference?

CPO removes the pluggable module entirely. LPO (linear-drive pluggable optics) keeps a pluggable module but removes its power-hungry DSP. AEC (active electrical cables) replaces short optical links with signal-boosted copper. CPO wins on power and density for the largest switches; LPO and AEC ($CRDO, $ALAB) compete for shorter reaches and more conservative deployments. They coexist today.

Is silicon photonics a good investment theme?

It's one of the most direct ways to play AI networking, but it's a component-supplier theme with real cyclicality and customer-concentration risk (a few hyperscalers and NVIDIA drive demand). The diversified approach is to track the layer — the Photonics / CPO basket — rather than betting on a single design win. This is not financial advice.

Related reading

  • The AI Supply Chain Map: Every Public Company in the AI Buildout — where photonics sits in the whole stack
  • What Is an External Laser Source (ELS)? — the light source CPO depends on, and why it sits outside the package
  • Sivers After the POET–Marvell Blow-Up: Why the Lasers Still Matter — the laser bottleneck inside CPO, in depth
  • Data Center Stocks: The Picks-and-Shovels Guide — the cooling and power layers CPO leans on
  • Photonics / CPO basket — live performance for the optical-interconnect names
  • 🎥 NVIDIA's New Photonic Technology Explained

What is co-packaged optics (CPO)?

Co-packaged optics integrates the optical engine — the part that converts electrical signals to laser light — directly onto the same package as a switch or accelerator chip, instead of housing it in a separate pluggable transceiver module. This shortens the electrical path, cuts power, and improves reliability, which is why it's being adopted for the highest-bandwidth AI networking switches.

Why does AI networking need CPO?

At 800G and 1.6T per port, pluggable transceivers consume too much power and fail too often to scale to clusters of hundreds of thousands of GPUs. The optical transceivers in a large AI fabric can burn megawatts on their own. CPO addresses both: NVIDIA cites roughly 3.5× better optical power efficiency and ~10× better network resiliency versus pluggables — power and uptime that translate directly into more usable GPUs.

What are the benefits of co-packaged optics?

Four step-changes versus pluggable transceivers, per NVIDIA: ~3.5× better optical power efficiency, ~63× better signal integrity (the ASIC-to-optics electrical loss drops from ~22 dB to ~4 dB), ~10× better network resiliency from 4× fewer, externally mounted lasers, and ~1.3× faster deployment. Broadcom's production data at Meta backs the power claim at roughly 65–70% optics power savings.

When do NVIDIA's CPO switches ship?

Quantum-X Photonics (InfiniBand) began shipping in early 2026 at low volume after slipping from late 2025. The first Spectrum-X Photonics Ethernet switches (SN6810, SN6800) started reaching select partners in June 2026, with the production ramp in the second half of 2026. Broadcom's CPO line has been shipping longer — Bailly since 2024–25 and Tomahawk 6 "Davisson" since October 2025.

Which companies make co-packaged optics?

The switch/integrator layer is led by NVIDIA ($NVDA), with Broadcom ($AVGO), Marvell ($MRVL), and Cisco ($CSCO) competing. The photonics are fabbed by foundries led by TSMC ($TSM), with GlobalFoundries ($GFS) and Tower ($TSEM). The optical engines, lasers, modulators, and assembly come from Coherent ($COHR), Lumentum ($LITE), Fabrinet ($FN), POET ($POET), Applied Optoelectronics ($AAOI), MACOM ($MTSI), and Lightwave Logic ($LWLG), plus private players like Ayar Labs. Fiber and connectors come from Corning ($GLW) and TE Connectivity ($TEL).

Why are CPO switches liquid-cooled?

Because they pack extreme power density — NVIDIA's Quantum-X Photonics is 3.95 kW in a 4U chassis — and co-packaging puts heat-sensitive optics next to the hottest silicon. Air cooling struggles to move that much heat while keeping the lasers in their stable temperature range, so NVIDIA's CPO switches ship with direct liquid cooling, extending the same loops the GPUs already use (Broadcom's lower-density 51.2T Bailly systems remain air-cooled). That makes cooling names like Vertiv ($VRT) downstream beneficiaries.

CPO vs LPO vs AEC — what's the difference?

CPO removes the pluggable module entirely. LPO (linear-drive pluggable optics) keeps a pluggable module but removes its power-hungry DSP. AEC (active electrical cables) replaces short optical links with signal-boosted copper. CPO wins on power and density for the largest switches; LPO and AEC ($CRDO, $ALAB) compete for shorter reaches and more conservative deployments. They coexist today.

Is silicon photonics a good investment theme?

It's one of the most direct ways to play AI networking, but it's a component-supplier theme with real cyclicality and customer-concentration risk (a few hyperscalers and NVIDIA drive demand). The diversified approach is to track the layer — the Photonics / CPO basket — rather than betting on a single design win. This is not financial advice.

Referenced on this page

  • Photonics / CPO basket
  • Silicon Photonics & Optical Interconnects
  • data-center power
  • nuclear
  • external light sources
  • Cooling basket
  • advanced packaging
  • Sivers deep dive
  • InP & Substrates basket
  • Networking / Retimers basket
  • layer two of the photonics stack
  • POET ($POET): The Photonic-Integrated CPO Engine Outlier
  • The AI Supply Chain Map