Marvell Technology Is Betting More Of Its Revenue On One End Market

Three-quarters of its sales already come from data centers, and the company’s own guidance widens that share even as it plans for cloud spending growth to cool.

Marvell Technology (MRVL) now draws roughly three-quarters of its sales from a single end market. In fiscal Q1 2027, the most recent quarter reported, the data center end market accounted for 76% of what the company sold, up from 74% in fiscal Q4 2026. On $8.7 billion of revenue over the trailing twelve months and a market value of about $162 billion, that share is the number a holder should be able to name. Not the growth rate, and not the margin: how much of the business answers to a single category of buyer.

Why The Guidance Pushes That 76% Higher

One quarter’s step from 74% to 76% proves nothing by itself. The guidance does. Management guides data center sales up about 50% in fiscal 2027, then about 55% in fiscal 2028, after a 46% increase in fiscal 2026. Nothing else in the company comes close to that pace: communications and other end markets, which produced $585 million of revenue in fiscal Q1 2027, are expected to grow about 10% in fiscal 2027 and at a low single-digit rate in fiscal 2028. If those figures land as given, the part of the business that could dilute the concentration is the part growing slowest, which makes 76% a floor for the mix rather than a high-water mark.

A Three-Quarter Share Set By A Handful Of Cloud Budgets

What makes a mix like that uncomfortable is who decides it. Management’s own account of where the spending lands is blunt: the top four U.S. hyperscalers spend the bulk of the capital budgets, so that is where the dollars go. On that telling, the largest buyers behind that 76% are a short list. And management is planning for cloud capital spending growth to moderate to around the 30% range in fiscal 2028 while planning for its own data center revenue to accelerate to about 55%, on the stated expectation that its interconnect business keeps outpacing that spending. The load-bearing assumption is therefore not that the build-out continues. It is that Marvell keeps taking a bigger share of spending that is growing more slowly than it did.

Does 18.5 Times Sales Leave Any Cushion?

All of this would matter less at a modest price. The stock trades around $183, about 58% of its 52-week high of $316.35, though it is still up about 140% over the trailing twelve months against 17.7% for the S&P 500. Even down that far from the high, it changes hands at 18.5 times sales, against a ten-year range running from 3.0 to 30.1. A price-to-sales multiple in the top decile of its own decade suggests the market is already paying for the guided ramp rather than waiting to see it, which is what turns a question about mix into a question about price. The underlying improvement is real: an operating margin of 16.4% over the trailing twelve months against a three-year average of 0.2%.

How Worried Should A Holder Be About That 76%?

Proportionately worried, not acutely. Concentration is not deterioration, and revenue grew 34% over the trailing twelve months. The narrower risk is that one variable, the capital budgets of a few cloud buyers, now sets three-quarters of the outcome, and at this multiple the price leaves little room if that variable disappoints. What would change the read is checkable when fiscal Q2 2027 is reported: whether data center revenue grows into the mid-40% range year over year guided for that quarter, on total revenue of about $2.7 billion, and whether the disclosed mix steps up again. If you are weighing a stock sitting this far under its own high, the fairer test is against the other names that have fallen this far and still have a case.

When One End Market Ends Up In Every Line You Own

The concentration inside a company is only half the question. The other half is what sits next to it. A semiconductor fund, a broad technology fund and a large-cap index fund can each end up leaning on the same handful of cloud capital budgets, so a portfolio can make one bet three times over without anyone deciding to. Owning a spread of businesses whose fortunes are not set by the same buyers is the point of the Trefis High Quality Portfolio, a rules-based basket of quality names rather than several versions of the same story. The Trefis High Quality (HQ) Portfolio has a track record of outpacing a benchmark that combines the three major indices – the S&P 500, S&P Mid-cap, and Russell 2000.