Top 6 AI Stocks To Buy Now (Biggest Bottlenecks)

Many investors are aware that artificial intelligence is likely to make investors very wealthy, but they're playing it wrong. See, the real idea is to actually bet on the bottlenecks. This is the scarce part of the system that could potentially slow everything else down because the supply is not nearly high enough to meet the very high demand. But this part gets a little bit trickier because the bottlenecks are always changing. So, in this video, I'm going to talk about the top six stocks that I'm identifying that could potentially be the next big bottlenecks for investors in the AI space. Starting with the number one biggest bottleneck is memory. And the stock that I'm looking at is Micron Technologies, ticker symbol MU. See, Micron, what they're end up facing is that they didn't build enough fabrications or factories to build enough memory to meet the amount of supply. And so customers are willing because the biggest models in the world are Opus Claude or Sonic Claude models that have put the highest amount of pressure on high bandwidth memory. Now Nvidia knows this and so they're ramping up the amount of DRAM or memory on their chips every single new architecture going all the way from the Hopper 100s all the way up to the rows of Fineman's. You're looking at roughly a 10x increase in the amount of memory per chip. But because of the lack of fabrications that Micron has available to them, they've been charging more and more higher average selling prices for DRAM going forward. And so even for this most recent quarter, which we saw today, Micron ended up announcing that they wanted to get to $33.5 billion in revenue at approximately an 81% gross margin. That's the highest gross margin you'll see across any hardware company in the world. And $19.15 of diluted EPS. Yet what they ended up seeing was 41.5 billion dollar in revenue skyrocketing past. Just take a look at this overall growth rate. Now over $350% year-over-year. Just last year they did 9 billion. This year they did $41 billion. The margin was not 81%. It was 84.5% and they did $35 billion of gross profit. But then on their earnings last year they did 1.8 billion. put this into perspective, okay? 1.8 billion and what you actually saw year-over-year was $28.2 billion. That is such an unbelievable increase. And their margins climbed to a new all-time high of over 70% margins on their overall net income. free cash flow ended up climbing up to a new company high of 17.5 billion which just quarter over quarter. So three months ago, which was also a company high of 5.5 billion, is more than 3x the amount that they just ended up showing off. Micron's cash has skyrocketed now up to $26 billion, which they can do a lot of. Maybe they give that back in terms of a dividend. Maybe they buy back shares. Maybe they get into new businesses. Right now, they're just paying off their debt. So they used to have $15 billion of debt last year. Now that's down to5 billion. Well, their cash has gone up to $26 billion. And one of the reasons why I'm getting so excited about Micron is because they're signing 3 to 5year long-term agreements with many of the biggest companies that are purchasing their memory. So this is not a flash in the pan as many people had originally thought because they are potentially creating sustainable earnings by having customers come back. And if they don't come back, they still have to pay tens of billions of dollars to Micron just to cancel the contracts to purchase these chips. And as they said in their market outlook, they expect supply demand conditions for both DRAM and NAND, the two different types of memory that they serve to remain tight beyond calendar 2027. So they expect this type of growth and the type of pricing power and high 80 plus% margins to start happening for years to come. Next quarter's guidance, they now expect to be at $50 billion, gross margins to improve to 86% and for diluted EPS to now be at $31 per share. Let's visualize that for a second. Take a look at this overall growth exploding over the last little bit. Margins continuing to climb up and much of that ended up going to the bottom line as they bring on $31 of EPS, which is up from last year at only $2. and let alone the fact that they absolutely dominated earnings last quarter. They expected $19 and they got 25. The quarter before that, they expected $9 and they got $12. What if this year ends up coming in at like 35 or $38 a share? Anything is possible right now whenever pricing power is in your favor like Microns is. The second bottleneck I want to talk about right now is not necessarily a primary bottleneck, but it's one that I think will build into one. So, this is quite early. We're talking about optics or Lumenum holdings, ticker symbol L I T. Lumenum supplies a lot of the photonics or co- package optics that end up going in between the chips that allow things like GPUs and CPUs or potentially to other racks to communicate amongst each other. And many of the biggest supplier of chips at rack scale like Broadcom and Nvidia are using more and more optics than ever before in their latest architectures. This has allowed Lumenum's growth to skyrocket in the last couple quarters, going from essentially 18% growth last year all the way up to 90% in just one year, bringing up that total growth up to $800 million in overall revenue. And this goes across both of their types of businesses. Both in their components revenue, which makes up for the large percentage at over 77% but $533 million, and on their systems level revenue, which is skyrocketing in its growth over 122% or about $275 million in revenue. Just like Micron however much earlier stage the gross margin is starting to skyrocket now at roughly 46 plus% and bringing on more and more profit on top of that accelerated level of growth. Net income is also rising and they just saw their first four quarters of straight profitability. And from the looks of it from their margins improving, it looks like they're going to keep that profitability going forward. Free cash flow is hitting all-time highs, and the amount of cash that they're printing is skyrocketing. Now at a new company high of $3.1 billion, growing by over 266% year-over-year. But I really want to stop here and talk about Lum's rule of 40. A rule of 40 is a old Wall Street metric that compares both the growth metrics that they have and their margins. So it doesn't matter if you put up 20% growth on 20% margins that would hit the rule of 40 or 30% growth at a 10% margin that would also be 40% on the rule of 40. Lmentum is putting up a combined growth rate in margin of now 126%. You are now in a very very rare category of companies that is putting up such unbelievable accelerated growth and this is up from 32% not even breaching the rule of 40 just one year ago. And next, I want to talk about their inventory because many people get this confused. There's some bears online that talk about Luminum's inventory and say, "Look, it's rising over time. This could be a potential problem because people are not buying what they have to offer." But that's not necessarily true. You have to end up breaking this down in terms of revenue to their overall inventories as a ratio. And what you end up seeing is that they're actually selling more and more of their inventory as a percentage of their overall revenue. So, of course, their inventory is growing because their overall revenue is growing, right? They're a bigger business now. So, of course, they're going to have more inventory. And that inventory is hopefully an early sign of what is to come, which is an accelerated growth rate into the hundreds of percentages, 105, 116. And these are just Wall Street projections. This is not even the company's own projections for which you're going to see this a lot during this video. Wall Street has been greatly incorrect in all of the companies that we're going to talk about today. greatly undershooting what they're going to actually put up in terms of growth and margins. Speaking of which, we need to talk about the most important bottleneck of all, compute and networking. Specifically, compute for this one. As Nvidia, ticker symbol NVDA, continues to sell more and more chips at higher and higher rates. Take a look at this. What we ended up putting up last quarter was essentially 85% growth on $81.6 billion in revenue. I mean, the scale of this company, they're the biggest company in the world, but the fact that they could reacelerate growth from 55% back up to 85% and potentially even next quarter doing over 100% growth, just assuming that they do 93.5 billion, which is almost right where Wall Street has it right now, would essentially bring them back to triple digits while being over a $5 trillion company. That has never been heard of. And yet whenever you take a look at their different types of revenue, this is all coming from data center spend. Companies like OpenAI, Enthropic, Meta, Microsoft, Amazon, all of these businesses cannot buy enough Nvidia chips because customer demand is extremely insatiable. At the end of the day, there's a lot of prices that are going up, whether that's optics or memory or many of the other bottlenecks that we'll talk about. But we cannot deny that Nvidia's pricing per chip has also skyrocketed. And take a look at this. Not only for the data center side, but especially the networking. Data center computing revenue of 60 billion was up 77% year-over-year, while data center networking revenue hit 15 billion, nearly tripling year-over-year. So, going back to that optical cabling and all this stuff that Lum is supplying, think about how much further we're going to continue to grow whenever networking revenue is tripling for Nvidia. There's aspects of this company that people are underestimating, but are growing extremely quickly. margins are improving, staying above 75% or bringing in $61 billion of gross profit. Well, also their operating expenses continue to fall. Research and development ended up falling. Still spending about 7.7%. So, they're not trying to go after new technologies. Of course, they are, but SGNA or sales general and administrative and stockbased compensation have now fallen to company lows. as they've continued to excel, they're giving more and more back to overall shareholders. By the way, that rule of 40 that Lmentum put up of 126%, Nvidia has a rule of 40 of 152% while doing it at a much much larger scale. Nvidia got their net income to $ 58 billion or a net income margin of 71% which also has to tie back to some of their marktomarket investments like investing in Intel back in August of 2025. less than a year ago, $5 billion investment has now returned them over $30 billion worth of gains. So in the last 37 weeks on average, they have taken a roughly $13.8 billion portfolio and turned it almost into $60 billion on the other end. That's just absolutely mind-blowing. But even their own business is doing well. Not only are they able to pick winners, but they're able to make a lot of profit. Look at their free cash flow. $48.6 billion. a new company high in next quarter they're expected to be the most profitable free cash flow business surpassing Apple next quarter and just like Lmentum that growing inventory has scared a lot of investors now growing almost to $26 billion people say that they not able to sell their products but whenever you actually break that down of what does their revenue look like comparatively to their inventories they're actually in the same steady pace that they've been over the last 3 to four quarters Nvidia just said this in their most recent quarter with analysts now forecasting hyperscaler capex to exceed $1 trillion in 2027 and Aentic AI beginning to proliferate all industries. AI infrastructure spending is now on track to reach 3 trillion to 4 trillion annually by the end of the decade. The end of the decade is like 3 years away and yet they believe that capex spend is going to go from $1 trillion and be on track to spend 3 to 4 trillion annually. That level of growth is unbelievable, but that flows down to the optical companies, to memory, to all of the other bottlenecks that we're going to talk about in the rest of this video as well. I want to continue on this bottleneck because I think it's the most important one. However, I want to look at it through a new lens through Broadcom, ticker symbol AVGO. We're going to be looking at more custom chips as Broadcom supplies the most powerful chips to both Google, like Google TPUs, and OpenAI's most recent chip that they just announced, which is unnamed right now. As Broadcom says here, with Google, we announced in April that we are entering into a long-term agreement to develop and supply multiple generations of their tensor processing units in AI networking. Our relationship continues to be a strategic and very substantial as we continue to deliver vastly superior technology in execution compared to other alternatives. That's what the CEO just said and it's showing up in their earnings as growth is reacelerating back up to nearly 48% and $22 billion worth of overall revenue. Most of this growth is coming from that AI semiconductor revenue side, but Broadcom is a very diversified business getting into many different areas of growth. Broadcom supplies a lot of the switches transistors Ethernet networking, all of the components that you need for AI and semiconductors and many of the other businesses. Broadcom's quite diversified there. Specifically on the semiconductor solution side, we're growing that rate to now 78.5% or $15 billion. And just like all the bottlenecks that we're going to talk about here today, the best way to spot a bottleneck pricing power. And you can find that within the margin. And so what you can see here is that growth, although we ended up growing by 78% is that the overall growth rate on their operating income, meaning after paying those expenses, 93%. which means you're having margin expansion. Customers are paying more and you can bring more to the bottom line. That shows up in their gross profit here now hitting over 69.5% delivering $15 billion of gross profit or consolidated net income at over 42% or roughly $9.3 billion worth of overall net income. Even Broadcom of all companies, super diversified, super early on in the custom ASIC game, is now at a rule of 40 of almost 117%. Another triple-digit rule of 40 company on this list. Just like the playbooks of before, you see them bringing up that inventory, but they're getting ready to ramp with OpenAI, and they're just getting started with TPUs as well for multiple years of architectures there. Total cash is climbing up over 107% growth year-over-year and their free cash flow is also climbing up to $10 billion or essentially a free cash flow margin of 46.25%. Broadcom is going to be the main competitor to Nvidia, but I don't want to have to pick which one is going to win at the end of the day. I want to pick both and so that's why I put them both on this list. For the next bottleneck, we have power. And there's many different ways to look at this, whether that's power or energy, but I wanted to talk about the actual power distribution. And for that we're talking about Vertive Holdings, ticker symbol VRT. Like I said, Verdive is actually distributing the power, potentially taking the power from the grid and then bringing that all the way directly into the chip. And there's many different processes and products that they use in between that that creates the Vertive ecosystem. Now, just like you're seeing in this beautiful video that they've created here, they talk about how they get the power, how they end up utilizing it, controlling it. By the way, there's multiple different ways that you end up doing this, not just for power, but for cooling, for thermal management, all of the aspects that doesn't necessarily include the direct GPU or custom chips. There's many things that Vertive actually needs and what a data center needs that Verdive also supplies. But this has allowed Vertive to increase their overall growth rates, now hitting roughly 30%, or bringing on roughly $2.65 billion worth of growth. that growth is diversified not just in North America but in Asia-Pacific in Europe, Middle East, African regions as well. And so that growth has been quite diversified and gross profit this quarter also neared all-time highs only just behind what we ended up seeing in the quarters before. But like you can see here, March is usually their lowest quarter. And so this is just the seasonality only to then bump up way, way higher the next quarter. So, we're likely to see a very high quarter like we see in guidance for Q2 for Verdiv. In net income, they're bringing in roughly about a 14% margin, which is much less than what we're seeing out of the other companies here. But this is a much more early stage bottleneck. Energy prices really haven't increased nearly as much as what people believe they will if Jensen is correct and the buildout continues in essentially 4xes over the next 3 plus years, which is what they talked about in their AI spending. If that continues to happen, these gigawatt plus data centers are extremely power intensive and so Verdive will be in extremely high demand as this puts a lot of stress on the grid. That will allow Vertive to have a lot of this pricing power, but they're still quite early. Yet, cash continues to climb to its all-time high, $2.5 billion, growing over 70% year-over-year. And their backlog, this is where you really see a lot of those future orders, is now at an all-time high, roughly growing at about 110% year-over-year. This is their future growth demand is skyrocketing right now. And free cash flow is also doing very well now hitting a very respectable $24.7% margin or bringing on roughly $650 million worth of free cash flow. And take a look at this. This is another metric that I just wanted to highlight about Verdiff as I'm learning more about this business. Their overall stockbased compensation is some of the lowest that I've ever seen in almost any industry at only 6%. This is extremely conservative to what other companies would be willing to pay their employees. Very small amount. They're definitely getting a lot of their employees via cash offerings, but that's good for shareholders if you expect that this company will be worth more in the future. For the next bottleneck, and this one cannot be undersold at all, is the actual data center. Now, for this company that I wanted to highlight is Coreweave, ticker symbol CRWV. This is the actual real estate construction and overall capacity development that you can get from not just coreweave but many of these other neocloud names and hyperscalers like Microsoft or Google or AWS. There's a bunch of these major companies that also do this. But this is a very important part to the bottleneck. Nvidia does not have an ability to sell chips if they don't have somewhere to put them. So the companies that are able to actually get access to that land have engineers that are able to set up the chips quickly. It's the fastest way for Nvidia's chips to be utilized or Broadcom or any of the other competitors. The data center companies are agnostic to which company actually has the best chips. Their bet is just on the idea that demand for artificial intelligence will continue to go up over time. This demand is not only skyrocketing in the hundreds of plus percentages in the case of Cororeweave here, but it's also recurring growth as people continue to pay their cloud bills. And look at Core Weef here for future demand. As Wall Street sees it, they're going to go from year-over-year from $2 billion up to $5 billion just year-over-year and then another 5.75 billion just the quarter after that. The growth here is just unbelievable. And Coree has constantly beat Wall Street expectations. So, it's likely to be even higher than what Wall Street is seeing right now. Now unlike Nvidia or some of the other companies that we talked about here today, Cororeweave is a very capital intensive business because they have to pay for all of the memory and the chips and all of the premiums that all the companies before us is charging. Coreweave is essentially paying that. They end up pushing that to the end consumer, but it is a very capital inensive business. So you need a lot of money. Margins here will end up showing that this is not a profitable business right now, but they are adjusted EBIDA positive. But as long as you end up seeing that backlog go from 14 billion up to 98.8 billion, they're going to continue to spend because customers continue to sign up for more over not one-year periods, but five 10 plus year periods. But yes, they are spending with most recent quarter showing capital expenditures now at 7.7 billion just in the last 3 months. But one of the biggest concerns with Core Wee or many of the NeoClouds is, hey, will those customers actually show up and pay their bills? But whenever your customers are the biggest and most affluent customers in the AI space, the likelihood that they're not going to pay you gets lower and lower. For example, Meta agrees to pay $21 billion to expand their AI infrastructure agreement. And yet, they're one of the most profitable companies in the world. They're likely to pay their bill alongside anthropic or open AI that continue to get more and more profitable every day. However, still a touchy subject right now. Now, one of the things that you need to highlight with Cororeweave and many of the other Neoclouds is how much in debt they're willing to go in order to obtain more chips and more of the actual components to find more customers. This is the part where it gets a little bit iffy. Unlike the other businesses that we've seen that have a ton of cash and not a lot of debt, Corweave is the opposite. They don't have a lot of cash and they're very strapped with debt. Whenever you look at the market cap of this business, you can see something along the lines like as of today $58 billion. But whenever you take a look at the enterprise value, that's their market cap minus cash because that would make the company cheaper plus debt, which would make it much more expensive, and you end up seeing that the real cost that you're paying for this business is somewhere along the lines of $90 billion. But yet more investors will be pouring into this company because they are now a part of the NASDAQ 100 and there will be forced purchases from passive investors. But maybe the debt will pay off considering the fact that there is a lot of different cloud companies. But there's only one cloud company according to semi analysis that is above the rest and that's coreweave. They have maintained platinum tier over the two different reports that semi analysis has put out called their cluster max ranking which essentially ranks all the different cloud companies and coreweave has come above nebius above Microsoft Azure above Amazon web services above Google cloud to become the number one spot which is platinum this is due to many different reasons whether that's great pricing the ability to set up new clusters extremely quickly have access to the best chips have the best software where there's a whole bunch of different reasons why, but at the end of the day, all you need to know is this is a wellperforming company that is in a very, very tight bottleneck. Now, for my next bottleneck, it's not necessarily a bottleneck, but it might be a solution to one of them. This is a power alternative, SpaceX Technologies, ticker symbol SPCX. Now, SpaceX might be the solution around one of the biggest bottlenecks, which is energy costs rising at the data center level. SpaceX has a whole mission in how they're going to eventually build up to this. But the total goal is to get data centers in space and use solar off an endless supply of our sun to essentially run these chips for free in the middle of space. Now, it would still probably use optical cabling and it would still probably use Broadcom or Nvidia chips in those data centers, but the real solution it's solving is power. Now, there's a whole other business to SpaceX like Starlink potentially giving customers access to broadband internet connections or directto cell connections to satellite. SpaceX also has XAI or Cursor X, the old Twitter and all these different things they could potentially offer. But the real potential bottleneck is solving one of them here and that's the solution to power. Now SpaceX is roughly twice the overall market cap of a micron. that is much much larger than the $18 billion that SpaceX does for revenue or the potential IBIDA of only $6 billion and Micron's printing roughly $28 billion a quarter in terms of their overall net income. I'm not investing in any way in SpaceX right now, maybe someday in the future, but I am invested today for Nvidia, Broadcom, Cororeweave, and Micron. But let me know in the comments which stocks you guys are most interested in and where is the biggest overall bottleneck. Make sure you subscribe if you want to see more content.