Picture a life without ChatGPT: you comb through dozens of research papers to find one good source for an essay, you write thousands of lines of code to build a barely functional website, and you rely on imperfect Google searches for your every answer.
Now, step back into today. AI is shaping our everyday life – how we work, shop, learn, and so much more. In the first quarter of 2026, 17.8% of the world’s working-age population used generative AI, including 31.3% of Americans (The Star). As the world becomes more reliant on AI, the physical components behind how AI works are no longer as sustainable. Thousands of chips talk to each other at once, and the copper wires connecting them are reaching their physical limits.
On 2 November 2026, Marvell Technology agreed to acquire Lumentum in an all-stock transaction to become a leader in AI connectivity, a market the companies size at $550 billion (Marvell and Lumentum). Lumentum builds optical circuit switches, lasers and transceivers that are meant to replace the traditional copper cables inside these massive data centers. The deal was welcomed because it pairs Marvell’s electrical chips with Lumentum’s optics, as copper is hitting its limits.
End of a Neutral Party
What came along with the deal was a loss of Lumentum’s neutrality. Before the deal, Lumentum sold optics to everyone: Nvidia, Google, Microsoft, and more. It made no chips of its own and wasn’t competing with any of them. However, Lumentum’s biggest customers are now buying optics from Marvell, a company that competes indirectly in the chip business with some of its customers (Gotrade). Because these customers were already buying from rivals like Broadcom and Coherent, they can now move their orders elsewhere and Marvell loses a part of what it paid for (Reuters).
It’s similar to how Apple originally bought its iPhone screens from Samsung. However, Apple has spent years adding other screen suppliers so that it doesn’t depend on a competitor. This is exactly what I believe Lumentum’s customers will do.
What Comes With the Deal
Lumentum’s top two customers made up 41.6% of its FY2026 revenue, and because Lumentum’s customers also buy from rival optics suppliers, it’s possible that they shift orders away and Marvell loses some of the revenue that it paid for. The structure of the deal also adds a second risk. Marvell agreed to pay a 23.9% premium in an all-stock purchase of Lumentum. Marvell’s shareholders will be diluted from owning 100% of the company to about 67%, while Lumentum’s shareholders are tied to Marvell’s stock through the 4.5 exchange ratio (Marvell and Lumentum). Marvell’s stock fell about 13% on 29 October, the day talks were first reported (Gotrade).
Furthermore, because the exchange ratio is fixed, Lumentum’s shareholders are exposed to Marvell’s stock. If customers start looking elsewhere for optics and Marvell’s stock continues to go down, Lumentum’s shareholders feel these effects before the deal even closes.
What Makes It Worth It?
Even though the deal was a strategic fit for Marvell, it’s dilutive to Marvell’s shareholders. An important factor to consider is how the deal could play out in the long term.
The questions that need to be answered are 1) does Marvell deliver the $1B+ a year in cost synergies (Marvell and Lumentum)? 2) does the optics market continue to grow? 3) do Lumentum’s customers stay? For example, if Lumentum’s products are hard to replace, or demand greatly exceeds supply, customers can’t afford to switch suppliers.
Furthermore, Lumentum’s largest customer went from 15.4% to 26.6% of revenue in one year (Lumentum 10-K), as the AI optics market was booming. Purchasing Lumentum in a bull market could mean that Marvell paid too much of a premium (~$20B). In addition, in all-stock deals, relative prices matter more than absolute ones.
It’s important to consider what Marvell’s stock is worth compared with Lumentum’s as of the signing. If Marvell’s stock was high, then it means that they gave up less in the acquisition.
My bet is that as the demand for AI infrastructure grows, this deal could justify its roughly $20 billion premium in the long run, but only if Lumentum’s customers keep buying from a supplier now owned by a company tied to their competitors.
The deal terms, the premium, the exchange ratio, the ownership split, the synergy target and the $550 billion market come from the companies’ joint announcement. The customer figures come from Lumentum’s annual report for fiscal 2026, which does not name the customers. Marvell’s fall on 29 October comes from Gotrade’s market report. The argument about Lumentum’s neutrality is mine.
The previous Trojan horse: The Hourglass →