Marvell to buy Lumentum for $115 billion in an all-stock bet on optical AI networking

Marvell Technology said on November 2, 2026 that it will acquire Lumentum, a maker of lasers and optical components, in an all-stock deal that values Lumentum at an implied $115 billion enterprise value. It is one of the largest deals yet in the AI hardware supply chain, and the market's first reaction was a sharp sell-off in both companies.

The strategic logic is easy to follow. The financial terms are unusual. And the timeline is long enough that the picture could look different by the time it closes.

The terms

From Marvell's press release: Lumentum stockholders will receive a fixed exchange ratio of 4.500 shares of Marvell common stock for each Lumentum share. That implies about $1,132 per Lumentum share and a premium of approximately 23.9% to each company's unaffected closing price on October 28, 2026, the last trading day before market speculation about a deal began.

After closing, current Marvell stockholders would own about 67% of the combined company on a fully diluted basis and Lumentum stockholders about 33%. There is no cash component, per CNBC. Both boards approved the agreement unanimously.

The companies say the combination would address a total market of $550 billion, offer "line of sight to more than $1 billion in annual run-rate cost synergies" and be accretive to Marvell's non-GAAP earnings per share in year one after close. All of those are forward-looking statements by the two parties.

The market reaction

Investors did not cheer. CNBC reported that Marvell shares fell 13% intraday on Monday and Lumentum fell nearly 19%. Before the sell-off Marvell had more than doubled over the past year and Lumentum had quadrupled, per CNBC, and Lumentum had more than doubled in 2026 alone.

A fixed-ratio, all-stock deal ties the price to Marvell's share price, so a falling Marvell lowers what Lumentum holders receive. That helps explain why Lumentum fell harder. We are not offering a view on either stock. The point is that the headline $115 billion is a snapshot at announcement, and it floats with the acquirer's price until the deal closes in the first half of 2028.

Why optics is the prize

Marvell's argument is that optics is moving from a component to the center of AI hardware design. Its release frames the issue directly: "rapidly growing bandwidth requirements and the physical limitations of copper-based connectivity are accelerating the transition toward optical technologies." As clusters grow from thousands to hundreds of thousands of chips, the cables between them have to carry more data over more distance on less power.

Lumentum matters because it makes the lasers. Nvidia's own materials name Lumentum as a supplier of lasers for its Spectrum-X co-packaged optics switches, per Optics & Photonics News. Marvell makes the electronics that talk to them: network switching, custom accelerators and the digital signal processors inside optical modules. The pitch is that designing the electrical and optical parts together, which the release calls "extreme co-design," gets more bandwidth for less power.

For the technical background, our separate explainer on co-packaged optics covers what the hardware actually does. For the wider capital cycle that is driving this demand, see our tracker of Big Tech AI capital spending.

What Murphy said

Marvell chief executive Matt Murphy told CNBC the combination would "create a new category of company here," that "the demand is exploding, and we don't see that changing anytime soon," and, on Nvidia's investments in the area, that "This is a double down on supporting Nvidia, who is the ecosystem leader in this area." Lumentum chief executive Michael Hurlston called it an "enormous" opportunity.

Two things are worth separating. Murphy's description of demand is an executive's view of his own market. The statement about Nvidia is more informative: it signals that the combined company's roadmap depends on the largest buyer's design choices. Concentration on one customer's architecture is a real risk for any supplier, and the release does not quantify it.

What could go wrong

The sell-off is the first test of the story. A premium of 23.9% over an unaffected price would normally lift the target, and instead the target fell nearly 19%. That suggests investors were more worried about the acquirer's share price, which sets the value of the offer, than excited about the price on paper.

The deal is a year and a half from closing, and several things can move in that time. The AI build-out could slow, which would hit optical demand directly. A merger of this size may draw antitrust scrutiny in several countries, although the sources we opened do not discuss regulatory review. Integration of two companies with different manufacturing bases, compound semiconductors on one side and CMOS on the other, is not trivial. And the all-stock structure means the final value depends on where Marvell trades.

Marvell closed a multibillion-dollar deal for the optical interconnect startup Celestial AI earlier in 2026, per CNBC, so this is a second large bet in the same direction within a year.

Our take

The direction is credible. Copper is hitting limits, optics is the answer the whole industry is converging on, and owning lasers, silicon photonics and switching silicon under one roof is a coherent plan. The numbers are not yet evidence. The $550 billion market, the $1 billion of synergies and the accretion claim are management projections, and the sell-off says investors want them tested.

What would change our mind: an independent market estimate that matches the $550 billion figure, disclosure of how much of the combined revenue depends on a single customer, and regulatory clearance without remedies. Until then, the safest reading is that this confirms optical networking has become strategic, not that the merged company will win it.