One of the biggest semiconductor sector winners last year was Marvell Technology(NASDAQ: MRVL), up 83%. Unsurprisingly, this was due to its burgeoning artificial intelligence chip business.
Marvell has a smaller but similar portfolio to Broadcom's (NASDAQ: AVGO) semiconductor business, which is a collection of ASICs (application specific integrated chips) that cloud giants use to create their own custom AI accelerators, networking chips and digital signal processors for AI data centers, and other less-attractive non-AI chip segments in broadband and telecom equipment, consumer electronics, and a small automotive chip business.
While the non-AI parts of Marvell's business have been in a down-cycle, its AI chip portfolio is booming, causing enthusiastic investors to bid Marvell up to a very high valuation. But does that make shares a buy, sell, or hold to start 2025?
The case for buying Marvell today
Like Broadcom, Marvell makes standardized IP blocks used by cloud giants to help speed up the design of their own custom ASICs. With Nvidia charging a high price for its GPUs, cloud giants are increasingly investing in their own custom semiconductors.
Broadcom appears to have a bigger ASIC business than Marvell, but Marvell has Amazon(NASDAQ: AMZN) as a customer. That has been a huge positive, as Amazon is the largest cloud computing platform and is making perhaps the biggest push to deploy its own homemade chips whenever possible. In fact, Amazon invested in AI start-up Anthropic, and as part of the deal, Anthropic agreed to train its newest large language models on Amazon's Trainium chips going forward, rather than Nvidia GPUs.
While customer concentration could be a concern for Marvell, last month Marvell and Amazon inked a comprehensive, multi-generation five-year collaboration agreement spanning ASICs, optical communications modules, ethernet, and PCIe equipment for AI data centers.
That long-term visibility bodes well for Marvell, as does its innovation engine. In December, Marvell also unveiled a new custom high-bandwidth memory (HBM) interface technology that will enable custom ASICs to process memory faster and with lower power. This is a crucial innovation, as HBM has become somewhat of a bottleneck to AI training and inferencing. So, the new technology further burnished Marvell's reputation as an AI innovator.
While last quarter's overall 7% growth for Marvell doesn't seem like much to write home about, the data center segment grew a whopping 98% even as its other segments declined. But with the data center business now making up 72.6% of revenue, Marvell has transformed from a diversified cyclical chip business to a high-growth AI franchise.
But valuation, as with many AI stocks, remains an issue
No doubt, the Marvell story is a good one, especially as cloud providers appear eager to ramp their custom ASIC footprint. However, a lot of good news appears priced into shares at this point. Currently Marvell trades at 43 times next year's earnings estimates, for the year ending in January 2026.
A high-looking multiple doesn't necessarily mean a stock is overvalued. So, investors will have to gauge the ultimate AI opportunity here.
Back in April, Marvell held its AI Day presentation in which management outlined its perspective on the opportunity ahead. In the presentation, Marvell saw its addressable market including ASICs, switching, interconnect, and storage controller chips rising from $21 billion last year to $75 billion by 2028. Management also saw its market share in these high-growth verticals growing from 10% to 20% by that time.
Assuming Marvell hits that target, it would make $15 billion in data center revenue by that time, up from the $4.4 billion run-rate today. Assuming the non-AI segments grow from about $3.3 billion last year to $4 billion by that time, Marvell would make about $19 billion in total revenue in 2028.
If Marvell maintains a roughly 30% adjusted operating margin, slightly higher than today, that would yield $5.7 billion in operating earnings and just under $5 billion in net income in 2028.
Currently, Marvell's market cap is over $100 billion, so investors are essentially paying over 20 times 2028 earnings in that scenario. That seems like a high multiple to pay for earnings four years out, especially with all the uncertainty abounding in the fast-moving space.
But it may be a reasonable price after all
One way in which Marvell's valuation may be justified is if its AI outlook from April has improved. After all, rival Broadcom just said in December that its own AI total addressable market had grown to between $60 billion and $90 billion just for its own current clients in 2027.
Marvell's April outlook appeared to take rivals like Broadcom into account, so it's possible the ASIC and AI networking outlook has improved since then.
So while Marvell may not be the best buy right now after its 2024 run, it may not be an outright sell, either. As long as the AI market continues to blossom, the stock appears a hold today.
However, if there is any inkling of a slowdown in the AI infrastructure buildout, investors should be prepared for the possibility of a hefty pullback. For those who don't already own shares, one should put Marvell on the watchlist, but perhaps wait for a better price before diving in.
Don’t miss this second chance at a potentially lucrative opportunity
Ever feel like you missed the boat in buying the most successful stocks? Then you’ll want to hear this.
On rare occasions, our expert team of analysts issues a “Double Down” stock recommendation for companies that they think are about to pop. If you’re worried you’ve already missed your chance to invest, now is the best time to buy before it’s too late. And the numbers speak for themselves:
Nvidia:if you invested $1,000 when we doubled down in 2009,you’d have $374,613!*
Apple: if you invested $1,000 when we doubled down in 2008, you’d have $46,088!*
Netflix: if you invested $1,000 when we doubled down in 2004, you’d have $475,143!*
Right now, we’re issuing “Double Down” alerts for three incredible companies, and there may not be another chance like this anytime soon.
John Mackey, former CEO of Whole Foods Market, an Amazon subsidiary, is a member of The Motley Fool's board of directors. Billy Duberstein and/or his clients have positions in Amazon and Broadcom. The Motley Fool has positions in and recommends Amazon and Nvidia. The Motley Fool recommends Broadcom and Marvell Technology. The Motley Fool has a disclosure policy.
