Airoha targets multibillion growth as 100G ramps, 200G tapes out

Airoha's 100G PAM4 DSP ramp is not a product milestone—it is the inflection point where optical DSP shifts from Broadcom's near-monopoly to a contested oligopoly. Technical cascade: Volume 100G-per-lane shipments will compress BOM costs at Innolight and Eoptolink, migrating the profit pool from silicon to system integration. The 200G tape-out deadline (end-2026) compresses the 400G/800G module DSP window to roughly 18 months, forcing CPO adoption earlier than hyperscalers originally planned. Discrete DSP cadence can no longer track AI cluster bandwidth demand exponentially; TSMC 5nm capacity becomes the hidden chokepoint. Compliance exposure: MediaTek's prior 5G baseband restrictions create a regulatory precedent. If 200G DSP lands on a 3nm node, export-control trigger probability escalates non-linearly. Airoha's Taiwan, China origin carries a structural discount in prevailing de-risking narratives. Competitive response: Broadcom's rational play is not price-matching but accelerating its CPO roadmap—changing the battlefield from chip to system. Marvell will likely lock in module partners vertically. Airoha's ecosystem window: 12–18 months. 24-month trajectory: 100G price war → 200G positioning → CPO architecture shift. Without 200G customer lock-in, Airoha's 100G scale advantage degrades into a one-time profit spike, not a structural moat.