저자: RickOrford · 외국인
요약 보기
브로드컴 1분기 FY26 실적이 컨센서스를 상회(매출 +29%, 반도체 +52%, AI 반도체 84억달러), 2분기 가이던스 약 220억달러(+47%), AI 칩 백로그 730억달러로 성장 지속을 강조. 선행 PER 40.8배는 높지만 PEG 0.77로 저평가 주장. 고객 집중도·맞춤칩 내재화(COT) 리스크 언급. 발언자는 AVGO 강력 매수 의견 유지.
다듬은 글
Broadcom just proved something that the market didn't want to admit. While investors keep questioning pretty much every AI stock that's out there, this company delivered the numbers. They raised the bar and may still be early in a much bigger run. So, what matters here is simple, right? The business is accelerating, the guidance got stronger, and as long as that keeps going, Broadcom may not look very expensive for very long.
My name is Rick Orford. I'm a Wall Street Journal best-selling author, been trading since 1999, and no, I'm not a financial advisor. That is a good thing. I break down the numbers so retail investors like us can make smarter, more confident decisions with our money. By the way, I wanted to give a shout-out to Seeking Alpha for the data in this video.
You can scan this QR code or click the link in the description for a free trial and a discount, and I'll even give you a link so you can follow along with the article that's published on their platform. Okay, so Broadcom reported its first quarter fiscal 2026 results on March 4th, 2026, and the numbers came in ahead of expectations.
Revenue outpaced Wall Street estimates by 0.9% and EPS on a non-GAAP basis beat by 1.32%. Since the report, though, Broadcom stock is up around 9%, and considering how the market's been well, questioning every AI logo that's trading today, I think that reaction says a lot, don't you? Because the story goes way beyond one single quarter.
It comes down to Broadcom's reported $73 billion AI chip backlog that they reported last quarter. For the first quarter, revenue rose 29% to about 19.31 billion, but it's where the growth that came from that actually matters. The semiconductor segment brought in about 12 and 1/2 billion, and that was up 52% year-over-year.
And it now makes up about 65% of Broadcom's total revenue. I mean, a year ago that number was 55% and I think that's important because chip revenue usually carries better margins at scale. On the other side, infrastructure software came in around 6.8 billion, but that only grew about 1% year-over-year.
But personally, I think that reflects stabilization after the VMware acquisition and that's fine. The whole point of buying VMware was to add steady high margin software revenue. I mean, it helps balance out the volatility that you see in the chip business. And when you look at the margins, the strategy starts to make even more sense.
The software segment alone carries a 93% gross margin where the semiconductor side is closer to 68. So, you'd think Broadcom made that bet for a reason and I say it looks like a smart one. At the same time, gross profit reached around 13.16 billion and gross margin expanded slightly to 68.1%. Operating income increased 37% to around 8.56 billion and R&D spending rose 32%, but this is actually worth paying attention to because Broadcom CEO Hock Tan has always been very tight on costs.
So, when a company spends an extra 700 million on R&D, it makes sense of demand was more than strong enough to justify the expense and that's exactly how I read it. As for the bottom line, net income rose 34% to around 7.35 billion and on the debt side, interest expense actually fell a little bit. It was 873 million and now it's 801 million and that happened as Broadcom refinanced part of the debt that it took on for the VMware acquisition.
Other income came in at 433 million. And after all of that, diluted EPS landed at a $1.50 a share. And that's actually up 32% from the same period last year. And it gets better. Cash from operations increased 35% year-over-year to 8.26 billion. And CapEx, well, it was only $250 million. So, [snorts]
free cash flow came in at over 8 billion. And and that's about 41% of revenue. Personally, I think that level of conversion is very impressive here, especially at nearly 20 billion in quarterly revenue. Broadcom also returned 10.9 billion to shareholders through dividends and buybacks. And all of that was funded from cash from operations.
Management also approved a quarterly dividend of 65 cents a share payable on March 31st, 2026. And they even authorized a new $10 billion share repurchase program through the end of 2026. But, of course, the real driver here is the second quarter guidance. Management guided revenue to come in around 22 billion.
And that would represent roughly 47% year-over-year growth. And adjusted EBITDA is expected to be about 68% of that revenue. In fact, I think the guidance is the big reason that the stock reacted so well after earnings. While she was expecting about 20.5 billion for the next quarter, and Broadcom came in meaningfully higher.
Now, before I go any further, picture this. If you like Broadcom right now, but you want a more disciplined way to potentially get into the stock at a lower price, one strategy you can consider is selling a cash secured put on Broadcom stock. That's a way to get paid while waiting for a better entry point.
And if you want to learn how that works, definitely check this video up here. Now, at the time of publication, Broadcom stock trades at about $346 a share. And that's up roughly 9% since earnings. And it's also up more than 87% over the past year. But year-to-date, well, let's just say I don't think it stays flat for very long.
Not especially after a report like this. Especially when the market reacted this positively. And what's interesting here is that Broadcom's near-term performance also goes against the broader rotation out of high multiple tech stocks that started in early 2026. I mean, that wasn't a Broadcom issue. It was happening across the group.
Yet Broadcom has clearly, well, they've bucked the trend. They've held up better than many of its semiconductor peers. Now, let's have a talk about valuation. Broadcom's forward PE, based on GAAP earnings, is 40.79 times. So, yes, on the surface, investors are paying about $41 for every dollar of future expected earnings.
And right now, that's roughly 40% above the sector median. So, at first glance, yes, Broadcom stock looks a bit expensive. But let's have a look at the PEG ratio, because I think this changes the conversation a little bit. I mean, PEG tells you how much investors are paying today relative to how fast earnings are growing.
And a value below one usually suggests the stock is undervalued. And Broadcom's PEG, or forward PEG, on a non-GAAP basis, is 0.77. And that's 40.47% below the sector median. So, hear me out here. The market may be pricing Broadcom's growth at a discount to what similar growth is getting elsewhere. And when you factor in the projected growth and the backlog, which I'm about to get into, well, a sub-one PEG very likely undervalues the earnings path.
So, what could push Broadcom stock even higher here? Well, I think the biggest growth driver over the next 12 to 14 months is Broadcom's AI accelerator business or it's XPU along with the networking chips that support it. If you look at the first quarter numbers, AI semiconductor revenue was up to 8.4 billion and that also beat management's own guidance.
During the first quarter, the CEO, Hock Tan, said the company now has a line of sight to AI chip revenue that would exceed a hundred billion dollars in 2027 and that's already backed by contract orders. It matters because this isn't just hopes and dreams. This is management that's pointing to demand that they already see.
In perspective, AI semiconductors now make up 44% of Broadcom's total revenue and a year ago that number was just 27%. That is a huge jump. And the growth is expected to speed up from here. I mean, for next quarter, Broadcom guided AI semiconductor revenue around 10.7 billion. That would represent 140% growth year-over-year and that estimate is also backed by actual demand.
Because Broadcom's AI related product backlog, well, it now exceeds 73 billion dollars and that covers about 18 months of deliverables. Now, during the earnings call, management did actually name some of the hyperscalers that are behind that backlog. Anthropic, for example, confirmed a custom chip order that's worth up to 21 billion dollars with deliveries from late 2026 and into 2027.
And don't forget, OpenAI was also confirmed as the sixth hyperscaler silicon customer and mass production of its first generation chips is expected by late 2026. At the same time, Alphabet's TPUs and Meta's MTIA's continue to scale. And more importantly, the CEO said that the supply chain is already secured.
Now, the real advantage in these relationships I see is being the co-design model. Broadcom works closely with each partner to custom-design chips for a specific workload, and then the customer licenses that design. And once the process is underway, switching mid-cycle would be painfully expensive. It would mean starting over or delaying execution by years.
And that's why I see it as a sticky moat for for Broadcom, and a very profitable one at that. I mean, if AI revenue hits 10.7 billion next quarter like management suggests, this sub-segment alone could contribute between 45 and 50 billion in annual revenue by the end of this year. And that creates a credible path to $100 billion in fiscal 2027.
And at a 68% adjusted EBITDA margin, I mean, that could imply an extra 13 to 15 billion in EBITDA from AI-driven growth alone over the next 12 months. Now, of course, no investment is without risk. And with Broadcom, I think the biggest immediate vulnerability here is customer concentration. Remember, a lot of this AI custom chip revenue is tied to just a handful of hyperscale customers.
That includes Alphabet, Meta, Anthropic, OpenAI, and Amazon. And together, they likely account for most of the $73 billion AI backlog. And separately, don't forget Apple still accounts for about 20% of total semiconductor revenue through its wireless components. So, yes, there is a bit of concentration here, and that matters because the nature of the chip model means that revenue from any one customer can swing depending on where they are in the design cycle.
I mean, when Alphabet ramps a new TPU generation, that alone can drive a billion dollars in quarterly sales. But between generations, well, that contribution can soften pretty fast. So, even though the revenue is sticky, it moves in waves. And during the earnings call, one analyst asked the CEO about the longer impact of customer-owned tooling or COT.
The concern is that some cloud companies are investing in internal chip design and may eventually handle more of this work themselves. And well, the CEO, he pushed back. And honestly, well, you'd expect him to. He said that Broadcom has been doing this for more than 20 years. And in generative AI, if you're trying to build your own chip, you can't afford a product that's just good enough.
And I think he's got a point. But I also think it's something worth monitoring. Of course, COT efforts are reportedly at least two generations behind Broadcom's current capacity. And that's according to the same analyst who raised the issue. So, I don't see it as an immediate threat. But there is also geopolitical risk.
The US Defense Secretary, for example, well, they recently designated Anthropic as a supply chain risk to national security. Now, that doesn't directly hit Broadcom specifically, but Anthropic has 21 billion dollars on order with Broadcom. So, if anything disrupts Anthropic's business trajectory, it could affect that commitment at least in some way.
And if just one of the top five AI customers pulls back, quarterly impact, well, it could easily be in the billions of dollars. And at a 68% margin, that would translate also into billions of dollars every quarter. Now, the $73 billion backlog does give Broadcom a cushion at least in the near term, but I do think it's smart to watch management's backlog commentary in the future quarters, especially for any changes in timing or customer mix.
Personally, I think the right way to frame this is through probabilities. The odds of all hyperscaler customers pulling back they're pretty much nonexistent. I mean, that's not a realistic scenario today. I think the more likely outcome is that we can get one customer hits a transition gap between chip generations, and that will create a temporary dip in Broadcom's numbers, not any structural breakdown in the business.
And that distinction it means a lot. So, does that make Broadcom a buy right now? Well, here's what I got to think, right? Broadcom delivered a very solid quarter. That makes a strong case even stronger. I mean, a backlog of more than 73 billion, a customer lineup that includes some of the biggest names in AI, and 41 cents of cash flow squeezed out of every revenue dollar.
That is hard to ignore. I think the market still sees Broadcom as a semiconductor company just having a good cycle. What I see is something a lot bigger. I see a company that's becoming the default design layer for hyperscalers. Granted, that kind of positioning doesn't fully show up in one or two quarters, but I do think the market is still in the early innings of pricing in what what that could actually mean.
And for that reason, I think my own strong buy rating on Broadcom stock remains in place. But now I want to turn it over to you. Do you think Broadcom is still overvalued? Do you think that the opportunity in AI is already priced in. Or if you're already holding Broadcom, well, let me know your price targets or or your biggest reason for continuing owning it.
Let me know your thoughts in the comments below. And while you're there, if you found this video helpful, don't forget to like and subscribe because it helps others find the video. It supports the channel and it makes sure that you don't miss out on my next deep dive. Well, that's it for me today, at least in this video. Thank you so much for watching and I'll see you next time.
Comments 0