Bandwidth is per-lane speed times lane count. Push speed and you hit the SerDes energy floor; push lane count and you hit fiber and connector cost. The new fact this year: the second wall arrived first.

850 nm is being pushed off the AI scale-up shortlist not because of speed but because of junction temperature. Moving to 1060 nm isn't swapping a die — fiber, PD, packaging and standards all have to change together.

The external laser source has gone from "one part inside a module" to a formal product category with specs, interoperability requirements and multiple competitors. Of the seven spec fields laid out at ECOC, power is just one; the real ticket is spec completeness and packaging throughput.

NPO is being treated as a long-term architecture not out of technical preference, but because of supply constraints. Separable architectures create component markets - and whether NPO or CPO wins, the laser stays outside the package.

What gets a supplier into the BOM isn't pJ/bit. It's fiber-attach takt time and a single speck of dust on an end face. This is a manufacturing race, not a photonics race.

After 1.6T there is no single winner: every reach has its own lowest-power answer. Two companies on this panel gave the same number — the reverse gearbox feeding slow-and-wide optics costs 4 pJ/bit, more than the optics themselves. And Oracle's field data pulled the spotlight back to the 4,096 fiber contacts on a front panel.

Today a rack draws 120 kW; the next generation will exceed 0.5 MW. At ECOC 2026, Microsoft asked a true customer's question: if optics were cheap enough, why keep this architecture? The answer came as a five-number spec sheet, plus a declaration that the die, not the package, is now the unit of design, assembly and failure.

An 850nm VCSEL that passed a 20-year, 100G production spec failed after 1,600 hours in scale-up rack thermal conditions, while a 1060nm device under identical conditions ran 5,000 hours unscathed. At ECOC 2026, Lumentum used one chart to show the VCSEL bottleneck was never speed—it is heat.

Chasing every option would force optical module makers to develop more than 50 SKUs. And at 1,024 lanes, OSFP beachfront density is just 0.07 T/mm versus a 0.3 T/mm target — a gap of over 4x. Huawei argues NPO means redesigning the whole switch, from sockets and thermals to PI/SI.

Microring DWDM has long been seen as the densest but hardest-to-control route. NVIDIA's ECOC 2026 paper takes "too hard to control" apart piece by piece: nine rings find nine lasers on a chip that can't see the spectrum, then re-shuffle live through an 80°C temperature drift, with even the forwarded-clock handover running at zero bit errors.

From 15 years at 10G to 4 years at 1.6T, the time for a data-rate generation to reach 10 million units has been cut again and again. At ECOC 2026, Lumentum laid out the numbers: half of module power goes to the DSP, only 24% of that DSP computes — and the real volume bottleneck is plugging in fibers.

Scoring three architectures on the same criteria, Ciena shows single-vendor CPO going from second-best to best on technology, but from efficient to at-risk on supply chain, because the ASIC and interconnect are sold as one unit. Open CPX's answer is not a better component but a socket that optics and copper share, and a spec written in under six months.

At 1.6T, choosing IMDD no longer buys you lower power — only shorter reach. With a reach-collapse chart and a standards matrix full of "Undefined," Nokia argues the real 1600G+ problem isn't physics but too many options and an industry slow to converge.

On the same day and at the same wavelength, Lumentum said OM3 and OM4 can't be used at 1060nm, while Berxel ran 100 meters of OM3. The number that really matters from this talk is ±22 µm: it downgrades optical packaging from precision alignment to simply placing the part.

At ECOC 2026, Huawei did something the industry rarely does in public: score NPO and CPO dimension by dimension. NPO won four, CPO two. But the real signal is the ranking, not the score - with supernodes heading to 4,096 accelerators in 2027, "can it be swapped when it fails" is put ahead of "how many pJ per bit it saves" for the first time.

OCS has left the lab, but today it is less a market than a customer list with one name on it. Cignal AI's ECOC 2026 numbers: ~95% of OCS ports go into Google's TPU network, and whether GPU scale-up becomes the next killer app depends on CPO going multi-wavelength — an 8x port-count gap that will split the supply chain.

A 409.6T CPO switch has 4,096 potential single points of failure, and about 80% of data-center network issues are cleaning-related. The EBO MSA's answer: no contact, no contamination. Ten of its 48 suppliers are Taiwanese.

Scale-across planning has shifted from "how many wavelengths do I need" to "how many fibers do I need", because regenerator sites can no longer supply the power to light that many fibers. Acacia mapped the whole optics landscape onto four scales in a seven-step chart, and the 400G/lane physics line is now setting the handover timeline for campus DCI.

At ECOC 2026, Coherent skipped the CPO sales pitch and brought a working machine: a 6.4T NPO with its own ELSFP at 4.7 pJ/bit — laser included. LightCounting puts the crossover at 2031. The CPO debate is over; what remains is a fight over the slope.

CPO's bottleneck isn't the chip — it's how 1,152 fibers get into 1U. At ECOC 2026, Corning broke CPO reliability down line by line: under 40 FIT in total, with 38 FIT coming from the fiber array alone. It is the first time CPO's volume-production risk has been translated into operating numbers — and the answer points to coupling.