저자: ParkevTatevosian,CFA · 외국인
요약 보기
AI 반도체 기업 브로드컴과 마벨을 매출, 영업이익률, ROIC, 밸류에이션으로 비교. 브로드컴은 수익성·ROIC 우위에 저평가(DCF 적정가 523 vs 현재가 353, 약 48% 상승 여력)로 제시, 마벨은 성장성은 높으나 고평가(DCF 207 vs 266, 약 22% 하락 여력)로 판단. 발언자는 브로드컴 보유, 마벨 비보유.
다듬은 글
Shares of AI companies are rising rapidly. Demand continues to outstrip supply, and the further we move, the more products are sold out for years to come. So it's understandable why many are interested in these AI companies, including Broadcom and Marvell, which are set to be among the main beneficiaries of this AI boom.
But it can be difficult to decide which stocks are best to buy. In this video, I'll try to help you make that decision by comparing these two companies on a number of important financial metrics, as well as evaluating their value to determine which stocks look more undervalued or better for long-term investment.
Let's get straight to the point. . Usually, the first thing I pay attention to when evaluating a company is sales. It all starts with sales. This is the first line in the income statement, because that's where it all starts. If you don't have sales, you don't have a business. From this perspective, Broadcom is showing phenomenal sales and phenomenal growth.
Its business has grown from about $35 billion in sales at the beginning of 2024 to about $90 billion in the past 12 months. This is partly due to the rapid growth in demand, and partly to the successful acquisition of VMware, which is successfully integrating into Broadcom's overall business. Broadcom helps companies like Alphabet create specialized integrated circuits, an alternative to buying products directly from Nvidia or AMD.
Companies like Alphabet began to appreciate this alternative to Nvidia and spent tens of billions of dollars investing and building more of these products. Meanwhile, Marvell operates in a similar category. They also deal with specialized integrated circuits. Marvell is taking greater advantage of data center connectivity by bringing all of these components together to work together in a large cluster to create value for customers.
Demand for Marvell's products and services is also growing rapidly, prompting Nvidia CEO Jensen Huang to suggest that Marvell could become the next trillion-dollar semiconductor company. Marvell's market cap is currently under $300 billion, so this would be a huge increase from current figures. Needless to say, demand for these companies' products and services is growing rapidly, and will likely continue to do so at least through this year, all of next year, and then starting in 2028 , probably in the first half of 2028 .
We have this visibility of demand for products and services in the AI ecosystem because these companies sign long-term contracts. So, we can see what the demand will be in 2027 and 2028 . After 2028, question marks begin to appear. I mentioned that I always start any analysis with sales, but right next to sales are profitability and unit economics —that's the order I look at them in my book.
A six-step system for evaluating public companies. I'm covering my process, and the first step in it is, of course, sales, which I called the customer value proposition . You make sales because you offer customers good value for their money or time. Besides, it is always easy to generate sales if you are willing to suffer huge losses in the end result.
Therefore, it is equally important to pay attention to unit economics or profitability. And in this regard, they both show significant improvement: Broadcom's operating margin increased to 48.6%. This is a significant increase compared to approximately 5% in 2017. Meanwhile, Marvell generates only a fraction of that operating margin, about a third—16.8%—but Marvell's operating margin is also growing.
As we saw earlier, Marvell is a much smaller business, so it is logical to assume that as it grows, its profitability may increase due to economies of scale. But now it accounts for only a fraction of Broadcom's profitability. At the same time, Broadcom's growth is more due to the semiconductor segment, which generates lower profitability than the software segment.
Nevertheless, profitability in this segment is sufficient to maintain the company's overall operating margin at or above current levels . Return on invested capital is the next indicator by which we will compare them. This indicator is important because it measures a company's performance relative to the amount of assets it invests in the business.
It favors companies that operate on a low- asset model and punishes those that use capital-intensive business models. The reason is that it is riskier. Capital-intensive business models are riskier because you are investing much more capital with unknown outcomes, especially in a new or unprecedented field like artificial intelligence.
Fortunately, Broadcom and Marvell use a relatively lightweight asset model. They outsource most of their production. Broadcom's return on invested capital (ROIC) has been impressive recently—25.75%. This is more than double the company's weighted average cost of capital. Meanwhile, Marvell's 12.61% figure barely reaches the company's weighted average cost of capital and is about half that of Broadcom.
So, now let's look at the cost estimate. So, up to this point, everything we've looked at in terms of performance, revenue, has significantly highlighted Broadcom. So it's surprising to me that Broadcom is selling at a cheaper valuation, but I can understand why that is the case. But let's see how different these estimates are.
Marvell trades at a forward price/ earnings multiple of 39. This is more than double Broadcom's forward price/earnings multiple of 18. In my opinion, the reason Marvell trades at a higher premium is because its growth prospects in the coming years are greater than Broadcom's. I mentioned that Marvell's market capitalization is less than $300 billion, while Broadcom's is over $ 1.6 trillion.
So, quite understandably, investors see less growth potential in Broadcom than in Marvell. And in the near term, because Marvell is starting from a lower base, its margins, revenues, return on invested capital, and cash flow are significantly lower than Broadcom's, so it has much more room for growth.
And that's why investors are giving Marvell a premium because its short-term growth prospects are better than Broadcom's. But does it deserve a premium of over 2x compared to Broadcom? I'm not entirely convinced of that. So I also wanted to look at the valuation using a discounted cash flow model that I created myself.
Looking at it this way, I come to a similar conclusion about Broadcom—it looks undervalued. I calculated the fair value at 523. The current market price is 353. So, I calculated a growth potential of about 48% over the next 12-18 months for Broadcom. Meanwhile, for Marvell, I calculated a fair value of 207.
The current market price is 266. So I calculated a potential downside of 22% over the next 12-18 months for Marvell. But of course, we'll learn results on September 30th. I will be going live to cover these results, and it will be available exclusively to channel members. Access channel membership by clicking the join button below.
So, I think the conclusion here is pretty simple. Comparing Marvell and Broadcom, Broadcom offers better business prospects historically, but looking to the future, Marvell offers better growth potential. However, given the difference in ratings, I would prefer Broadcom over Marvell. And that's exactly the choice I made for my portfolio.
I own Broadcom stock. I do not own Marvell stock. I am not interested in buying Marvell stock at any prices close to current levels. I prefer Broadcom stock, as well as several other semiconductor companies, to Marvell.







Comments 0