Broadcom’s Selloff Just Created A Massive Opportunity!

[00:00:00] Speaker 1: Hey guys, welcome back. Last week, Broadcom, the second most important AI chip company on earth, dropped nearly 15% in a single trading session. Media outlets everywhere were blasting the same exact headline. Broadcom plunges after revenue misses and it caused a systemic sector flush and it caused retail anxiety and panic across the board. But here's the thing. If you pull up the official press release on Broadcom's investor relation page, you uncover a glaring distortion because this infrastructure giant delivered an exceptional quarter. Yet a tiny third-party mathematical rounding error of just about 0.35% was weaponized to trigger a massive programmatic market sell-off. Last week was not an operational failure. It was a mechanical market overreaction. In today's video, we're going to discuss and dismantle exactly how the market completely misinterpreted Broadcom's print. I'm going to discuss also how the AI infrastructure bottleneck is evolving right now. I'm going to show you a historic keynote speech that came out of Computex last week that gave me the ultimate high conviction playbook and strengthened my conviction overall in one particular layer of the AI infrastructure. And finally, at the end of the video, I'm going to show you a massive signal that just hit the newswire earlier today that just validates the entire thesis that we talk about in this channel. So make sure you stay to the end. Make sure you pick up that massive signal. But before we get started, as I always say, make sure you guys hit like, make sure you subscribe to the channel. We have seen massive growth over the last few videos. And I just sincerely wanted to say thank you to each and every one of you that have either been on this journey on this channel since the beginning or joined somewhere along the way. I truly appreciate all of your engagement, all of your kind of just support of the channel, how you engage with it, drop questions, like videos. It has been an awesome experience. It's an awesome community that we are building together. So let's continue to expand it together. If you haven't subscribed yet, do me a favor, hit the button now. All right, so let's step back. And the thing that I want to talk about in this video is how the flush that we saw last week is potentially setting up an absolute gift for us as long term investors. And I think I had the fortunate, unfortunate opportunity yesterday to kind of step back away from the noise and take a look at everything that occurred because I was stuck in a massive amount of cancellations and delays trying to get out of Chicago yesterday. It was awful, guys. I got home probably after 1 a.m. this morning. It was a rough travel day, but it's great to be back in Texas. But like I mentioned, I had this multi-hour window to just look back over last week, look at the performance of some of the key companies that we track and we talk about in this channel. But I also had the chance to look at all of the keynote presentations that came out of Computex last week. So I'm talking about Nvidia, Qualcomm, Intel, ARM, as well as Marvell. Those last two were really impressive videos. And I'm going to drop the links to both ARM and Marvell's keynote presentations at Computex because I highly encourage each and every one of you, take the 45 minutes to an hour for each of those. Just listen to their CEOs and the roadmap that they lay out because it's really impressive. I walked away with even more conviction in these two companies, even more conviction in what they're trying to solve and how these two companies will help break through the bottlenecks that we see in the market overall. But let's step back and let me tell you why my conviction was strengthened overall. Because we're seeing a massive movement towards agentic AI and that essentially was the theme, the entire theme of Computex. We're looking at these autonomous personal agents that is going, they're going to completely break the legacy hardware architecture that we have today. When software transitions from simple chatbots to distributed reasoning agents that actively browse the web, execute complex multi-step plans and constantly pull long-term memory, the demand for compute high bandwidth memory and ultra low latency connectivity escalates exponentially. And that'll bring me right back to the absurdity that we saw last week and what we saw after Broadcom's earnings, the misunderstanding of their earnings and how the market interpreted. Because that's one thing I wanted to dig in and look at during my multi-hour break at O'Hare on Saturday or yesterday. So let's dive into the facts. Let's look at actually what Broadcom printed and how the market interpreted it overall. So let me pull up an article first that was posted on CNBC, let me pull that up for you because I saw so many of these when I was looking back through the tape. And this is part of the story how the, if you just read the headlines, particularly the key points below, you can misinterpret it. And I think we see some of these programmatic algorithm bots that just scrape headlines and they trigger selling overall 'cause this to me is part of the story. You can see the headline Broadcom stock plunges on weak software sales, unchanged AI chip forecast for the year, but look at that first key point. Broadcom reported weaker expected revenue in its fiscal second quarter earnings report on Wednesday. No clarifying, no qualifying that statement at all. You can interpret that as that they missed their target and that couldn't be further from the truth actually. Broadcom exceeded their guidance and their target. They were slightly off the whisper numbers, the consensus numbers from the analysts overall. But let me show you what they actually printed. But first, let's take a look at what it did to the stock. So here's Broadcom. We know what they did on Friday down seven, you know, almost 8%, but look at the one week. It said closing in on 14%. I said nearly 15% right there, but they had a massive kind of slide last week overall. But look, let's look what they actually printed because here's what they printed. Revenue was 22 point, we'll call it 22.2 billion up 48% from prior year period. Their earnings per share was just over $2.44. So that was over estimates. I think their, their estimate there was $2 and 40 cents. So they're about 4 cents overall. And if you scroll down even further, you can see that their semiconductor revenue business was, was exceptionally strong. So their revenue from AI down here in the bottom paragraph was 10.8 billion. It was up 143% year over year. And it was driven by increasing demand for custom AI accelerators and AI networking, according to their CEO. This momentum is going to carry forward into the next quarter as well, we're there. They expect semiconductor revenue from AI to grow over 200% year over year to $16 billion. They also set a massive guidance moving forward. So in Q3, we expect consolidated revenue growth to increase 84% year over year to almost $29 and a half billion. You know, with non gap operating margin stable at 67% reflecting our strong operating leverage. So guys, that is not, that is not a weak earnings that actually beat their earnings to look at they beat their guidance from last quarter. I want to show you that's a 22 point call, 2 billion. Here's what they guided for in Q1 for this quarter. They actually guided for 22 billion flat with a 47% increase. So they beat it by almost 200, almost $200 million overall. But the challenge was, is that they didn't beat the consensus number that got put out the whisper number. As I mentioned, that number was set higher. That number was set about 22 point almost 22.27 billion was the whisper number. And the fact that they didn't hit that is what caused this massive sell off that but think about it. This is an infrastructure giant that even when you look at that whisper number, they were only off by $80 million. That's a huge loss. On a mega cap business like Broadcom has, in the immense scale of their business overall, this is a computational rounding error of just 0.35%, as I mentioned in the beginning. But this was the critical piece that the news wires didn't hit is when Haktan opened up for his earnings for his presentation, he explicitly said that the total revenue reached 22.2 billion up 48% above our guidance. So think about the sheer mechanical distortion that was happening here. A company like them, a core infrastructure giant like Broadcom, they set internal targets, they execute flawlessly in the quarter, they beat their own guidance. They print a massive net income increase of 88% year over year to almost to $9.3 billion. And they drop a blistering Q3 revenue forecast of almost $29.5 billion, completely wiping the floor with Wall Street's expectations. So why did the stock flush 15%? It's because they missed this arbitrary aggregated mathematical mean from a third party analyst model. And they missed it by a third of 1%. So algorithmic headline scraping bots, as I mentioned, they essentially triggered a massive panicky sell-off. This was not an operational failure from Broadcom at all. This was, it was a mechanical market overreaction to what we saw from their print. But here's the beauty for us as long-term investors. This overreaction has created a massive structural mispricing of so many strong assets in the market right now. Mispricing of so many assets across the entire sector. Everything is essentially on sale right now, which puts us back in the seat of power, let's say, as investors, where we can get assets that are significantly undervalued in the market right now. But I do want to jump and just kind of shift gears. While the short-term retail crowd was panicking over Broadcom's phantom revenue miss, anyone tracking the actual physics of data center deployment was glued to the stage of Computex last week. Because Marvell CEO, Matt Murphy, he walked up to the podium onto the stage and articulated a profound structural shift that explains exactly why Broadcom's dip was a gift for all of us investors. Murphy posed a fundamental question. The question was, what defines the performance of AI infrastructure? Historically, investors focused entirely on the processor, the GPU, the XPU, or the leading edge foundry node. But let me show you what Matt's perspective was. Let me pull up that video or we can walk through it together. [00:10:34] Speaker 2: But that's still not the defining characteristic of the system. Because one processor, no matter how fast it is, no matter how much memory it has attached to it, is simply not enough for today's AI workloads. You need tens of thousands and eventually millions of processors working together as a single massive compute engine. compute engine. That's why computing at this scale is fundamentally a connectivity challenge. And increasingly, it is the architecture and characteristics of connectivity that defines the performance of the system. [00:11:10] Speaker 1: As Matt stated, one processor, no matter how fast, no matter how much memory it has attached to it, is simply not enough for today's AI workloads. We're going to see tens of thousands and eventually millions of processors working together as a single massive compute engine. That's why computing at this scale is fundamentally a connectivity challenge. We have officially arrived at the phase three or the third bottleneck, the connectivity wall. That's the way Matt's describing it right now. If the memory wall created Micron's trillion dollar opportunity, the connectivity wall may create the next generation, the next trillion dollar companies in the infrastructure winners. That's why Jensen called Marvell the next trillion dollar company when he took the stage. Connectivity is no longer a secondary support component. It is an absolute defining architectural characteristic that dictates the performance of the entire AI system. To solve this connectivity challenge, Marvell has spent the last decade executing a massive, deliberate corporate transformation. Ten years ago, less than 10% of Marvell's revenue came from inside or just came from the data center. Today, over 75% of Marvell's revenue comes directly from data infrastructure. And they are the undisputed global leader in data movement. And the core of their thesis hinges on a terrifying physical constraint that the copper wall has now moved inside the server rack. Let's listen to Matt and what his comments are directly on this. So let me pull him up and we can listen to his comments together. [00:12:41] Speaker 2: It's going to move again and it's going to take over the rack itself. So this is creating an explosion in demand for the optical industry. Incredibly complex engineering challenges are coming and along the way. So why is this happening? So it's not just somebody's preference to go do this. This is physics. The distance a signal can travel over a copper cable is inversely proportional to the bandwidth. So every time you double the bandwidth, you have to cut the distance in half. Today, the highest speed production systems in the world run at 200 gigabits per second per lane, just to give you an example. So with that bandwidth, the cable length is limited to roughly two and a half meters. Now, by comparison, systems running at 100 gig could use about five meter cables. And the height of the rack is about two meters. So once you account for all the routing inside the rack, two and a half meters is right at the limit. So when we move to 400 gig, we can no longer fully connect the rack with copper. So the wall is moving and it's moving now. And going forward, even the connections within the rack will become optical and the whole industry knows this is coming. So we've been preparing for this moment, not just Marvell, but the industry. And you see this in Taiwan, by the way, in the supply chain and the ramp up that's happening. The ramifications for this are actually enormous because each time the wall moves one step to the right, the number of connections that you have goes up by at least an order of magnitude. So it's creating this explosion in demand, as I mentioned, and the optical supply chain needs to scale up massively and be ready. [00:14:15] Speaker 1: So as Matt mentions in the video, when the industry transitions to next generation 400 gigabits per second signaling, it becomes physically impossible to fully connect an AI server rack using copper. The electrical resistance generates catastrophic thermal loads and drains massive amounts of power just to move electrons through metal. The industry is hitting the copper ceiling. The electrical backplates must be replaced with optical fiber, moving data via photons of light through glass. And unlike electrons and copper, the distance light travels through glass is completely independent of bandwidth. It brings us to Marvell's explosive new product portfolio that they showcased at Computex as well. It's all around co-packaged optics. Marvell showed off an absolute engineering masterpiece on stage at Computex. Because in a traditional switch, the silicon sits at the center of the motherboard and the copper traces all the way to the front panel where pluggable optical modules are inserted. And at high ultra frequencies, those motherboard traces degrade the signal. But in Marvell's co-packaged optic architecture, they can completely eliminate the copper motherboard traces. They take their 51.2 terabit switch silicon and heterogeneously bond 16 distinct 3.2 terabit optical engines directly onto the exact same substrate packaging. Light paths plug directly into the chip housing itself. This architecture, it slashes latency down to a sub 10 nanoseconds and it eliminates massive system power requirements and breaks the physical boundaries of the server. Let me show you what he showcased on stage overall. So let me pull that up and we'll take a look at it together. [00:15:54] Speaker 2: Over here, you have a traditional ethernet switch. This is our 100T teralink switch that we announced today. And you guys are the first to see it actually, everybody here in the room. You can see the switch in the middle of the board. Copper traces inside the PCB carry the signal to the front panel, which is here. And this is where all the optical modules plug in. Now, let's move over here. This is a CPO based switch right here. Now, notice that there's still the switch silicon in the middle. That's right in the center of the die of the package. In this case, this is our 51.2 T switch. And all around the edges are 16 3.2 T optical engines. So the 16 times 3.2, you get 51.2. So the fiber is directly attached now to these engines. It's not to the front panel. So we've completely eliminated the copper traces on the PCB. Light comes directly out of the package. Okay. This is a very, very complex piece of engineering. And it was very cool to be able to show this off today. [00:16:55] Speaker 1: So that is a very cool piece of engineering, as you mentioned. It's exactly the reason why they have the unmatched connectivity portfolio. You know, inventions, innovations like that overall. And it also explains why the king of AI himself, Jensen Wong, made a surprise visit during Matt's keynote. NVIDIA recently executed a massive structural partnership with Marvell, focusing on their NVLink Fusion. Integrating Marvell scale up networking tightly into NVIDIA's Blackwell and future Vera Rubin architecture. And when Jensen stands on a global stage, points to his partner and calls them the next trillion dollar company, he isn't doing that out of politeness. He's doing it because NVIDIA physically cannot scale their multi-million GPU clusters without Marvell's optical interconnect technology. And I mentioned at the beginning that I want to show you a massive signal that verifies our entire thesis. Here's the signal that I wanted to show you. This was just published today on Tom's hardware. You can see Google signs 920 million monthly compute deal with SpaceX. Let me scroll down and show you the detail right here that we just pulled out. So, SpaceX just announced that it closed a multi-year deal to provide compute capacity to Google. The agreement is worth 920 million per month and will begin in October of this year. It's expected to continue through June of 2029. Reuters said the transaction includes, listen for this, 110,000 NVIDIA GPUs plus CPUs, memory, and other components needed for AI processing. So, think about the macro reality here. Google is arguably the largest single owner of AI compute on Earth with their custom TPU pipeline. They are expanding their 2026 CapEx forecast to over 180 billion. Yet, they are cutting an almost billion dollar monthly check to rent 110,000 NVIDIA GPUs, CPUs, and memory from SpaceX because of the demand for the Gemini Enterprise. It is completely outstripping what they can supply internally. When a trillion dollar hyperscaler has to rent physical infrastructure from a rocket company for bridge capacity, it tells you that the compute memory and connectivity layers aren't just strong. They are locked in a structural supply squeeze. So, this brings us back to the question that I had in the beginning. So, as investors, as long-term investors, the nearly 15% drop we witnessed in Broadcam last week. Guys, in my opinion, it is a gift wrapped in financial noise. The news wires panicked over that arbitrary 0.35% third-party consensus model variance, completely missing the fact that Broadcom beat its own operational guidance and reported that AI hardware bookings are backlogged through the roof. When macro headline bots, when they create artificial panic, they temporarily drag down the highest quality assets in the market. The core pillars of the AI infrastructure stack, particularly ones we talked about today, the connectivity tool, the tollboots of Broadcom and Marvell, they just went on sale, in my opinion. Maintain your conviction, look past the clickbait headlines, and focus on the architecture. I hope this video helped. I hope bringing you some of this content kind of helped strengthen your conviction like it did for mine. I hope it brought you some more clarity as well, because if it did, hit the like button, hit the subscribe to the channel, turn on the notifications so you don't miss any more of our market signal updates we drop on each and every day. I'll see you guys in the next video. Have a great Sunday night.