Divesting What Doesn't FitNarrow moat

The counterpart to buying is selling, and in August 2025 Marvell completed the sale of its automotive Ethernet business to Infineon for $2.5 billion in cash¹. The business was expected to contribute $225-250 million of revenue that fiscal year — so the price represented roughly ten times revenue for a franchise Marvell judged non-core.

The logic is focus. Automotive silicon is a good business with long design cycles and a different customer set, and competing there meant spending engineering capacity that the data center could absorb entirely. Selling it converted a distraction into $2.5 billion of cash at an excellent price, and simplified the story the company tells.

It also has an accounting consequence worth remembering: the gain flatters fiscal 2026's reported results, which is why GAAP earnings per share that year exceeded the non-GAAP figure and then collapsed the following quarter. Watch what Marvell does with the proceeds, and watch for further divestitures of the legacy lines. A company narrowing toward its best market is usually a good sign; one selling assets to fund a stretched balance sheet is not, and the two look similar from outside. Marvell generated about $1.4 billion of free cash flow in fiscal 2026², so this sale was strategic rather than forced.