Marvell Bets $115 Billion That AI's Future Runs on Light

Marvell Bets $115 Billion That AI's Future Runs on Light

Everyone wants to talk about the GPUs. Meanwhile, the unglamorous job of moving data between those GPUs has turned into one of the hottest businesses in tech, and Marvell just put about $115 billion on it.

4.5 Shares, One Very Big Laser Company

On November 2, Marvell Technology announced an all-stock deal to acquire Lumentum at an implied enterprise value of roughly $115 billion. Lumentum shareholders get 4.5 Marvell shares per share, about $1,132 each, which Yahoo Finance puts at a roughly 23.9% premium to October 28 closing prices. Once the deal closes, Marvell holders will own about 67% of the combined company and Lumentum holders about 33%.

Lumentum makes optical components and systems, and its pump lasers and optical circuit switches are driving most of its growth. Business is booming: fiscal Q1 2027 revenue came in at $1.26 billion, up from $533.8 million a year earlier. Both boards approved the deal unanimously, and the companies expect more than $1 billion in annual run-rate cost savings. Marvell CEO Matt Murphy stays on as chairman and CEO, and three Lumentum directors, including CEO Michael Hurlston, join Marvell's board.

Plumbing Turns Out to Be Glamorous

AI data centers have a traffic problem. Training and serving big models means shuttling enormous amounts of data between chips, racks and buildings, and copper wiring runs out of room. Optics, meaning lasers, switches and the components around them, is how the industry keeps scaling, and Marvell wants to own more of that stack from silicon to light.

The detail worth noticing is the timeline: the companies expect to close in the first half of 2028, pending shareholder and regulatory approval. That's a long time in AI years. Marvell is betting that the hunger for AI connectivity is a decade-long buildout, not a bubble, and it's paying for that conviction in stock today for a payoff that's years away. Regulators also get a long look at a combination this big.

For everyone downstream, the infrastructure layer is consolidating fast. The cloud and AI services your business runs on are increasingly built on a few very large suppliers, and that affects pricing, capacity and how quickly new capabilities show up.

Chips get the headlines, but this deal is a reminder that AI is also a networking problem, and it's very expensive to fix.

The giants are rebuilding the pipes under AI, and we help businesses plan the part they control, from where your data lives to how your site and apps connect to AI services, so tell us what you're building.

Source: Yahoo Finance