On November 2, Marvell and Lumentum officially announced an all-stock acquisition agreement with an implied enterprise value of approximately $115 billion, or about $1,132 per share. Upon completion of the transaction, existing Marvell shareholders will hold approximately 67% of the combined company, while Lumentum shareholders will hold approximately 33%.

By equity value, this ranks among the largest technology M&A deals globally in 2026, far surpassing Marvell's 2021 record of acquiring Inphi for $10 billion. Why did Marvell have to buy—and why now?

Lumentum's financial trajectory answers that question on its own. According to publicly disclosed data, Lumentum's FY2026 full-year revenue reached $3.014 billion, up 83.2% year-over-year; fourth-quarter revenue hit $1.01 billion, surging 109.3% year-over-year, with adjusted gross margin reaching 50.4%.

Lumentum management has stated explicitly that optical component capacity has been sold out through early 2029, with roughly 70% of demand for some products still unmet next year, and new capacity typically requiring three to five years to build. Nvidia's $2 billion strategic investment in March 2026 to lock in Lumentum's laser capacity is itself a pricing signal for this scarcity.

Marvell's optical interconnect business is accelerating as well, with FY2026 revenue of $8.195 billion, up 42% year-over-year. Data center revenue accounted for 79% of the latest quarter, and the company expects optical interconnect business growth to exceed 70% by the end of FY2027—a significant raise from its previous 50% guidance.

Yet Marvell's optical DSP chips must work in concert with Lumentum's lasers and optical transceivers within the same optical module, and Marvell itself was previously a purchaser of Lumentum's laser technology. Against a backdrop of persistently tight supply and lead times measured in years, locking in upstream light sources through acquisition is essentially buying an "entry ticket" for its own optical interconnect growth.

Marvell CEO Matt Murphy called the deal a "once-in-a-generation" opportunity amid the "fundamental architectural shift of AI infrastructure toward optical connectivity." The combined company's serviceable market expands to $55 billion, with expected annual cost synergies exceeding $1 billion, and the deal is expected to be accretive to adjusted EPS in the first full year after completion.

The transaction comes at a time when the optical interconnect market is growing systematically faster than data center capital expenditure. LightCounting's publicly available forecasts show that optical transceiver and related product sales reached $23.8 billion in 2025, up 55% year-over-year, with 60% growth expected to continue in 2026. By 2031, the global market is expected to approach $60 billion, with a compound annual growth rate exceeding 20% from 2025 to 2031.

Overall, the deal's implication for the optical industry is that the competitive dimension of optical interconnect is being redefined. In the past, optical component manufacturers' valuations were anchored to module shipment volumes and per-product gross margins. But when light source capacity becomes a hard constraint on AI cluster expansion—and when optoelectronic co-design determines system-level connection efficiency—manufacturers with light source capacity and optoelectronic co-design capabilities gain the opportunity for repricing.

The transaction is expected to close in the first half of 2028, subject to approval by both companies' shareholders and regulatory clearance. Looking ahead, what truly deserves continued attention is whether optoelectronic co-design will become standard practice in the next round of optical interconnect competition.