This week the market read SemiAnalysis's CPO-delay report as a death knell for optical components. It read it backwards. If CPO ramps a year later, pluggable optical modules — and the upstream EML lasers already sold out through 2027 — get an extra year of earnings. Demand hasn't weakened; only the oversold share prices have.

This week the bull case for optical communications got a new spokesperson: Broadcom's income statement instead of analyst decks. Demand is now locked in by earnings, but what's truly scarce can't be bought with earnings: the hardest bottleneck in the chain is a laser in short supply through 2027. The more real the demand, the more valuable the light source.

The Big Four CSPs lifted 2026 capex to USD 725B, NVIDIA declared the arrival of agentic AI, and Cisco doubled its AI order outlook. Yet Chinese module makers' "revenue up, profit lagging" signal shows margin in this optical bull run shifting upstream to EML/InP and epitaxy.

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.

Goldman Sachs raised its optical module market forecast and sent Taiwanese suppliers limit-up three days running, while Eoptolink's earnings call pinned CPO/NPO volume production on 2027. Which year's story is the market pricing in early?

CoW hybrid bonding yield is first gated by the initial bonding-front angle. By tuning the angle to D so the bonding wave advances in order from center to edge, ASE lifted yield on 8×12×0.05mm dies from 2.8% to 99%, and held 95–99% on ultra-thin dies with 102.4µm warpage, 6µm pitch and organic interfaces, greatly widening the process window.

Edgecore admitted 1.6T LPO is "not yet successful," so the stopgap is gone and CPO can no longer wait a generation. Three MSAs at OFC 2026 push CPO toward swappable engines, while SONiC only now has the models it needs. Hardware decides who ships; software decides who deploys at scale.

On the surface, Microsoft's keynote showcased in-house silicon; underneath, it publicly moved the AI infrastructure bottleneck from accelerators to power, cooling and networking. For Taiwan's supply chain, Microsoft's own "70% of components come from Taiwan" is an entry ticket, not a moat: the money is flowing to system-level co-design.

Elaser's 2Q26 EPS tripled QoQ, but the real signal is its revenue mix: AI optics jumped from 32.7% to 45.5% in seven months. COSA is the cash cow, AOC the second engine, and its biggest bet, ELSFP, is still not in volume production. STT flags the three numbers to watch next quarter.

In one week, TSMC moved CPO from slideware onto a production schedule, and Lumentum seized pricing power at the light source. Two narrow gates, TSMC COUPE capacity and 200G EML supply, now decide who gets to play in the 2026 1.6T cycle.

Taiwan's optical stocks hit limit-up on Oct 2, but the Senate bill and the reported FCC route are two different paths. The 3.2T start line and the 65% US-BOM threshold decide who wins orders, and when.