0:00 While everyone's been glued to the Iran war, there's been a ton of news that investors can't afford to miss. So, in this video, I'll catch you up on eight major stories that are already moving markets and quietly changing which stocks are about to win big. Your time is valuable, so let's get right into it. The first story is the Iran war itself and its impact on the stock market. CNN's Fear and Greed Index has been stuck in extreme fear for the last four weeks, one of the longest stretches in recent history. It's easy to understand why everyone's so scared. But it's much harder to turn it to your advantage in the stock market. Just this week, President Trump warned that a whole civilization will die tonight if Iran didn't back down. That kind of language obviously makes normal people want to
0:45 pull their money out of the market. Then just one day later, we got a twoe ceasefire. Oil prices fell by 15% and stocks ripped higher as investors breathed a collective sigh of relief. If you want to be greedy when others are being fearful, CNN's fear and greed index is a great way to do just that. This index blends seven important indicators like stock price momentum and breath, market volatility, safe haven demand for bonds, and the ratio of put to call options. And then it calculates a single score from 0 to 100. 100 is delusional greed and zero is total widespread panic. The market has been in extreme fear for the last month. And now that you know how to track that, let's talk about the biggest stories actually worth getting greedy over. While everyone was focused on Iran, Nvidia
1:32 announced $2 billion investments in Nebus and Marll technologies. I'll cover Nebius first and Marll a little later in the video. Nebius, ticker symbol NBIS, is a Neocloud company focused on providing AI infrastructure for sovereign AI and highly regulated industries. Think finance, healthcare, and defense. And they just got the ultimate stamp of approval from two of the biggest AI companies on Earth, Nvidia and Meta Platforms. On March 11th, Nvidia announced that they'll invest $2 billion into Nebius as part of a larger long-term partnership to deploy more than 5 gawatts of Nvidia systems by 2030. 5 gawatt is enough power to run almost 4 million homes. As part of the agreement, Nebius will get early access
2:17 to Nvidia's newest Vera Rubin platform. A few days later, Nebius announced a 5-year AI infrastructure deal with Meta Platforms worth up to $27 billion. As part of this agreement, Nebus will deploy 12 billion of compute capacity to Meta by 2027, including some of the first large-scale deployments of those Vera Rubin GPUs. Meta has an option to buy up to $15 billion more. If Nebius doesn't sell that capacity to other customers first, analysts estimate that once all this infrastructure is fully built out and online, it could add anywhere from 4 to6 billion in annual recurring revenue for Nebus, which would be around a 300 to 400% increase from their current $1.25 billion run rate in
3:03 2025. Separately, Nebius's management is guiding for 7 to9 billion in total ARR by the end of 2026, which implies more than a 5x increase year-over-year. As a result, discounted cash flow models like Simply Wall Street's calculate the fair value of Nebia stock to be over $600 per share, which makes it around 80% undervalued today. Set another way, Nebia stock would have to 5x from today's price to reach its fair value after these two massive AI infrastructure deals. But Nebius is still a small and volatile stock that trades at a high price toearnings ratio. So it may not be for everyone, but this next stock definitely is. Microsoft has been trading at its 200E moving average
3:51 for the first time in over 13 years. That 200E moving average is the average closing price of a stock over the last four years of trading days. This is a great metric to watch because it smooths out the short-term noise and reveals a stock's long-term trend. For strong companies, it often acts like the floor where big patient buyers step in. So, when a quality stock like Microsoft falls down to that line, it signals a rare long-term buying opportunity, not just another random dip. In fact, this is exactly the kind of setup that the late great Charlie Munger loved. He famously said that if all you ever did was buy highquality stocks at the 200E moving average, you would beat the S&P 500 by a wide margin over time.
4:37 Microsoft stock is down by over 20% year-to date, making it the second worst performer of the Magnificent 7 right after Tesla. The difference is, of course, that while Tesla's revenues, gross profits, and earnings per share have all been in decline, Microsoft has grown all three by about 15% in a single year. And from a much larger base, DCF models like Simply Wall Street's calculate Microsoft's fair value to be around $450 per share, with some analyst price targets over $500. That means Microsoft is anywhere from 17 to 25% undervalued, giving it anywhere from 20 to 33% upside from today's prices. By the way, if you've ever wondered whether a stock is actually worth its market price, you are not alone. Finding a
5:23 company's fair value is one of the hardest parts of investing, which is why I use Simply Wall Street. Their stock screener isn't just about filtering by PE ratios or growth rates. You can actually screen for companies that look undervalued based on different approaches like discounted cash flow models, analyst estimates, and many other built-in checks so that I can choose the right metrics for the right companies instead of forcing everything into one model. Then I take those ideas and move them into watch lists for each company. I can choose which valuation I care about most. DCF models for the hyperscalers, analyst estimates for smaller companies, and even other researchers with their own price targets. and then I can sort that whole list by most undervalued. That turns my watch list into a ranked valuedriven
6:08 list that I can actually act on. I'll share my favorite stock screener settings and watch list in the description below. Simply Wall Street is jam-packed with tools that I actually use because they save me time and they help me buy the best stocks at the best prices. And now you can get the best price on Simply Wall Street by using my link below. 40% off until April 13th and then 30% off after. Talk about a great investment. And speaking of great investments, Nvidia also invested $2 billion into Marll technology, ticker symbol MRVL. On March 31st, Nvidia announced a strategic partnership to plug Marll directly into their ecosystem via Envy Link Fusion. NVLink Fusion is a plug-and-play connection system that lets custom chips from other companies
6:53 snap into server systems and talk to Nvidia's GPUs at full speed. That means Nvidia can still make money on infrastructure like CPUs and networking gear, even in data centers built around another company's chips like Intel's CPUs, Google's TPUs, or even AMD's GPUs. In this case, Marll will build custom AI chips called XPUs and high-speed networking gear that will connect to Nvidia's ecosystem through NVLink, while Nvidia provides the rest of the rack. The Vera CPUs, the Bluefield DPUs or data processing units, ConnectX network cards, SpectrumX switches, and so on. Nvidia and Marll are also teaming up on silicon photonix and optical interconnects, a topic that I recently made an entire video on. This comes
7:38 right after Marll closed its $3.25 billion acquisition of Celestial AI, which makes photonic fabric, a technology that uses light instead of copper to move data between chips and memory inside AI data centers. So, one of the leading custom chip designers with cuttingedge optical technology is getting pulled into Nvidia's inner circle with a multi-billion dollar endorsement. But there's another shift happening inside data centers that many AI investors missed because it's on the CPU side. After over 35 years of only licensing designs to other companies, ARM recently announced their first ever self-branded chip called the AGI CPU. This chip can pack up to 136 cores per CPU and 64 CPUs per rack and it's
8:24 designed specifically for AI inference in data centers. Meta Platforms is the first customer and ARM says that OpenAI, Cerrus, and Cloudflare are also lined up to use it. But I wouldn't rush to buy ARM stock just yet. In my personal opinion, ARM making their own AI chip puts them in the awkward position of competing with some of their biggest customers like Qualcomm, Amazon, Microsoft, and even Nvidia. All of which pay for a license to use ARMs architecture. Also, chips take years to go from design to volume production. So, we'll have to see how performance compares to the competition, including ARM's own customers by the time it actually starts shipping. The real takeaway for investors here is that ARM is directly entering the AI chip race, which could reshape the entire data
9:11 center CPU market over the long term. All right, let's talk about something that happened right as I was writing this video. Anthropic is the company behind the Quad family of models like Quad Opus 4.6, Quad Co-work, Quad Code, and so on. On April 6th, Anthropic announced a new compute deal with Google and Broadcom that gives them access to roughly 3.5 gawatts of next generation TPU capacity starting in 2027. That's on top of the 1 gawatt of Google's TPUs that Anthropic already has coming online this year. That combined 4.5 gawatt is roughly enough electricity to power over 3 million homes or roughly every household in the greater Chicago metro area. Broadcom, ticker symbol AVGO, is
9:57 the one designing the custom TPUs and the networking equipment connecting them all together under a long-term supply agreement with Google that goes through 2030 that effectively locks in all that revenue over the next few years. And that makes them a great picks and shovels play positioned right between the hyperscalers and the biggest AI model companies on the planet. Anthropic dropped another number that's worth paying attention to. Their revenue run rate is already over $30 billion a year, up from around 9 billion at the end of 2025. So in less than three years, Anthropic has gone from basically pre-revenue to out earning most of the companies in the S&P 500. One way for retail investors to get exposure to Anthropic is through the Fundrise Innovation Fund, ticker symbol VCX. VCX
10:44 began trading publicly on March 19th, so investors that were focused on the Iran War may have missed the fund's debut. But here's where things get even more interesting. Anthropic has a new, still unreleased model called Claude Mythos, which is said to be much more powerful than their current Frontier models. In fact, Mythos is so powerful that Anthropic won't release it to the public. For example, it's already discovered thousands of zeroday bugs and exploits across every major operating system and web browser, including a Linux exploit that could give attackers full control over a machine. Since Claude Mythos could be so dangerous in the wrong hands, Anthropic formed a coalition called Project Glasswing with companies like Crowdstrike, Palo Alto Networks, Broadcom, Amazon Web Services,
11:30 Cisco, Microsoft, Nvidia, and JP Morgan Chase to use Mythos defensively. The idea is to let these partners point Mythos at their own infrastructure and code bases to find and patch critical vulnerabilities before criminals or other nations could exploit them. In my opinion, there are two big takeaways here for investors. First, Anthropic is no longer just another open AI competitor. They're quickly becoming a core infrastructure and security partner to some of the biggest enterprises on the planet. And second, cyber security companies like Crowdstrike, ticker symbol CRWD, and Palo Alto Networks, ticker symbol PNW, are getting early access to some of the most powerful tools that their competitors don't have yet, giving them a huge head start
12:16 against the next wave of AI powered cyber threats. All right, another big breakthrough happened while everyone was focused on Iran. This time in memory and data storage. On March 25th, Google DeepMind unveiled a new compression technique called Turboquant, which directly attacks one of the biggest bottlenecks in running large AI models, the KV cache, which is a big part of an AI models memory footprint. Basically, TurboQuant can shrink the memory needed for the KV cache by around sixfold with no losses in model accuracy. And some benchmarks show that it speeds up key inference calculations on GPUs by up to eight times. Within hours of the announcement, memory stocks were dropping. Samsung, SKH Highix, and Micron all fell by 5 to 7% and SanDisk
13:04 plunged by double digits on the worry that AI data centers might need far fewer memory chips than previously expected. If you've been watching this channel for a while, then you remember what happened with DeepSeek. Nvidia stock dropped because Deepseek R1 showed that using a mixture of smaller expert models that only activate when they're needed could drastically reduce compute costs. But when costs go down, total demand goes up by much more since many more people can now afford it. So after DeepSec reduced compute costs, demand for Nvidia's chips skyrocketed instead of going down. That's called Jevans paradox. And I believe the same thing is going to happen with memory after DeepMind's Turbo Quant breakthrough. Right after that, on March 18th, Micron reported earnings for what might be their best quarter they've ever had.
13:51 Revenues jumped by around 75% quarterover-arter and 196% year-over-year while earnings exploded by over 700% thanks to the insane demand for high bandwidth memory and management guided for $33.5 billion in revenue for next quarter which would be another 40% growth in just 90 days. Like I've been saying for years now, memory is no longer a commodity. It's a critical component for AI model performance, latency, throughput, power costs, and ultimately how smart the model can feel. And I'm no longer the only one who thinks so. On April 2nd, Roundhill launched the first ever dedicated memory ETF, ticker symbol DR AM. This fund is
14:38 fully focused on memory names with Samsung, SKH Highix, and Micron making up around 70% of the fund and names like SanDisk, Seagate, and Western Digital rounding out the rest. Roundhill is explicitly pitching memory as one of the most constrained layers of the AI stack. It's the part that's hardest to scale, which makes it potentially the most profitable part as AI workloads keep growing. So check out the DRAM ETF if you want to hold the entire memory market in a single ticker or if you want direct exposure to SKHENX. They just announced plans to pursue a US listing in the second half of 2026. So make sure to keep SKH high stock on your radar as well. Either way, I expect memory demand to explode after TurboQuant, not go
15:24 down, especially as AI models get bigger, context windows expand, and more applications move into production. So whether you're invested in Micron, DRAMM or EWY, which is iShar's South Korea ETF that's also heavily invested in Samsung and SKHEX, I think you'll be pretty happy with your investments over the long term. But we can't talk about direct listings without talking about SpaceX, which filed for what could become the biggest IPO in market history. SpaceX is expected to go public this summer at a rumored $ 1.5 to2 trillion valuation, making it worth significantly more than companies like TSMC, Broadcom, and Meta Platforms. Now, Elon Musk himself went on X and called the number BS. So, take that range with
16:11 a grain of salt. But either way, this is about much more than reusable rockets. Earlier this year, SpaceX completed the largest private merger in history when it absorbed XAI, the company behind the Gro chatbot and family of models in an all stock deal that valued the combined company at $1.25 trillion. And in case you didn't know, XAI formally acquired X this time last year. So, the actual entity going public includes reusable rockets, Starlink satellite internet, the Grock AI models, and a social media platform, all under one ticker. Personally, I hope this goes public with the ticker symbol SXX. That would be pretty on brand. But either way, the big long-term goal here is orbital data centers, which bypass energy and real
16:58 estate constraints here on Earth by launching compute into space, powering it with solar panels, and then beaming the results back to Earth via Starlink. That's not science fiction. That's the actual strategic road map that SpaceX filed with the FCC. And to make it all happen, Musk announced Terrafab on March 21st, a 20 to25 billion AI chip manufacturing complex in Austin, Texas, that's designed to produce custom silicon for Tesla's autonomous vehicles and humanoid robots, SpaceX's satellite systems, and those future orbital data centers I just mentioned. The goal here is to produce up to 1 terowatt of compute capacity per year. And to make that happened, Intel formally signed on as Terafab's manufacturing partner. Intel will be contributing the chip
17:44 design, the fabrication, and the packaging capabilities. Just to be clear, a terowatt is a 1000 gawatt, which is about 50 times more than the 20 gawatt of chips that the entire AI supply chain on Earth produces today. One funny way that Elon and Intel could hit that goal is by making really, really inefficient, power- hungry chips. Wait, so maybe Intel is the perfect partner after all. All jokes aside, hopefully this video helped you catch up on some of the major market stories that got buried by headlines of the Iran War. And if you feel I've earned it, consider hitting the like button, subscribing to the channel, and even sharing this video with someone who might find it valuable. That really helps me out, and it lets me know to make more market recaps like this. And tell me in the comments if you
18:29 enjoy this style of rapid fire news and how I could make it even more valuable for you. Either way, thanks for watching and until next time, this is Tickerol U. My name is Alex, reminding you that the best investment you can make is in you.
