If you thought that AI chips had reached the high water mark, then you are entirely wrong.
Nvidia has been one of the only games in town, and that is a strong sign of a first-mover advantage.
In fact, the ecosystem could benefit if several chip companies could rise to the occasion to infuse that extra bit of supply.
Nvidia is on record, saying they can’t meet demand.
Well, we have finally reached the next phase of the AI chip story, and that is the next company stepping up to the plate.
AMD (AMD) has been working furiously to get into the AI GPU game, and it appears as if their harvest is just around the corner.
For the past 18 months, Nvidia (NVDA) has dominated the GPU industry with a ball-busting market share of up to 98%.
Nvidia's H100 GPU set the benchmark for AI training and AI inference.
The H100 is still a sizzling product today, and Nvidia continues to struggle with supply constraints because demand is so high from leading AI companies like OpenAI, Amazon, Microsoft, and more.
Those supply challenges have opened the door for competitors like Advanced Micro Devices to swoop out of nowhere. The company announced its own data center GPU called the MI300X at the end of 2023, which was specifically designed to compete with the H100. So far, it has lured in some of Nvidia's top customers, including Microsoft, Oracle, and Meta Platforms.
AMD forecasts the MI300 series will propel its GPU revenue to a record $4.5 billion in 2024 - an estimate that has already been raised twice.
Nvidia still is the champion - it started shipping its new H200 GPU earlier this year, which is capable of performing AI inference at nearly twice the speed of the H100.
Nvidia is now focused on its latest Blackwell chip architecture, which paves the way for the biggest leap in performance so far. The new GB200 NVL72 system is capable of performing AI inference at a whopping 30x the pace of the equivalent H100 system.
AMD is preparing to ship another new GPU next year called the MI350X, offering a staggering leap in performance of 35x compared to CDNA 3 chips like the original MI300X.
Advanced Micro Devices has explicitly said the MI350X will compete directly with Nvidia's Blackwell chips.
Nvidia plans to ramp up shipments of Blackwell GPUs during its fiscal 2025 fourth quarter.
A 114% increase year over year in data center revenue is what it looks like on the balance sheet for AMD.
Developing artificial intelligence (AI) software wouldn't be possible without data centers and the powerful graphics processing chips (GPUs) inside them.
This is where we stand – at the beginning of an AI-induced supercycle in technology stocks.
AMD is clearly the 2nd horse in the race that will pick up market share on Nvidia.
This could easily turn into a duopoly of GPU chip companies, and readers would be ignorant to not apply this knowledge a trading regimen.
Wait for a substantial dip to buy into AMD shares.
You’ll regret it if you don’t, especially long-term.
Unshackling the restraints on human labor – that is where tech is headed.
I’m talking about AI.
Robots aren’t able to perform complicated tasks and that is the holy grail of AI.
If headway is made just on this one issue then the sky is the limit.
Profits are then unlimited and the world will change into something we could have never imagined.
If stakes weren’t high enough, the next explosive leg up in tech shares is now centered on this concept.
There is only so much balance sheet maneuvering can add to the bottom line.
Magnificent 7 stocks who are experts are juicing up the balance sheet will gradually run out of levers to pull.
Technology stocks demand that management move the needle along because the alternative is that the company will get left behind.
When the Department of Defense commenced its robotics challenge in 2015, the stated goal was to develop ground robots that can aid in disaster recovery with the help of human operators.
Nearly a decade later, generative AI is accelerating that learning curve, pushing human-like machines to pick up new tasks in real-time.
And just recently, Tesla (TSLA) presented an updated version of its Optimus robot at Tesla’s Investor Day and showed it roaming a factory floor. CEO Elon Musk touted the robot’s potential, saying it had the ability to push the company’s market cap to $25 trillion.
Humanoids that can adapt to existing environments have long been seen as the ultimate test if they can work alongside humans in spaces built for them.
Nvidia (NVDA) is driving rapid development through an ecosystem built specifically for humanoids. It combines high-powered chips that process data at high speeds with a digital world that allows users to train robots on skills applied in the real world.
Just this past summer, Nvidia unveiled “NIM Microservices,” a visual training ground that allows generative AI models to visually interpret their surroundings in 3D.
Nvidia’s ecosystem now enables robots to train using text and speech input, in addition to live demonstrations.
Humanoids have already begun taking their first steps into reality. Musk has said two Optimus robots are working at Tesla’s Fremont factory, and he expects a few thousand to be deployed by next year. Amazon (AMZN) has partnered with Oregon-based Agility to utilize its Digit robot at a test facility. Apptronik is working with Mercedes-Benz to integrate Apollo into its manufacturing line.
The goal is to adapt humanoid for the future which will allow them to operate beyond industrial use. They could become as ubiquitous if companies are able to scale and bring costs down to $10,000 per machine.
Technology is still in the stage of calculating how they bring the expenses under control.
It is not very cost-effective if a company needs to spend 5 times the actual cost of running the AI division on retrofitting the environment for a humanoid and resetting the language models for different tasks.
Much of these technical aspects are being worked out, and these companies are inching their way closer to a day when companies might be able to work fully without a human worker or alongside a minimum amount of workers.
Tesla is a company long-term that needs to be looked at and this assumption is solely based on their robotics and humanoid business. It is highly plausible that Elon Musk is at peace with sacrificing his EV business in the medium time as long as moving up the value chain to become the leader of what is next which is looking more like robotics using AI.
Musk is skating to where the puck is next and that is where the future will be.
https://madhedgefundtrader.com/wp-content/uploads/2019/05/cropped-mad-hedge-logo-transparent-192x192_f9578834168ba24df3eb53916a12c882.png00april@madhedgefundtrader.comhttps://madhedgefundtrader.com/wp-content/uploads/2019/05/cropped-mad-hedge-logo-transparent-192x192_f9578834168ba24df3eb53916a12c882.pngapril@madhedgefundtrader.com2024-10-04 14:02:442024-10-04 15:46:30Tech Wants To Get Rid Of Labor Costs
I have been pounding on the table urging my readers to buy chip stocks.
Why?
Because chip stocks will carry the Nasdaq to higher highs.
Jump on the bandwagon while you can.
My thesis was validated when Micron stock (MU) jumped over 17% yesterday and is up over 20% for the week.
That type of stock appreciation isn’t as widely found in the tech sector anymore now that much of the tech sector is deadweight.
The sub-sector that isn’t dead weight is chips and specifically the AI chips which Micron is part of.
So when we talk about growth, you won’t hear stuff like earnings or revenue growing in the single digits.
We hear numbers more similar to revenue growing at 90% or 100% or even 300% in some cases.
The outperformance in growth is helping these stocks reach greater heights and this is just the beginning.
The commentary has been widespread that AI data spend on chips is going through the roof.
Micron’s management told us they raised guidance because of a more favorable pricing environment as well as robust demand for Micron's memory chips used in data centers to power artificial intelligence.
Executives now expect the market for high-bandwidth memory (HBM) chips used in AI data centers to increase to $25 billion in 2025, up from $5 billion this year — and heightened demand for its HBM chips to bring in multiple billions of dollars next year.
Micron is the first chipmaker to report quarterly results this earnings season and their stellar earnings bode well for the rest of its peers.
The company reported revenue of $7.75 billion — 93% higher than last year.
Micron distinguishes itself by partnering with, rather than competing against, industry superpower Nvidia (NVDA). Micron supplies memory chips for Nvidia’s hotly demanded GPUs.
The company is also set to benefit from a bill awaiting signature from President Joe Biden that would loosen environmental requirements for microchip projects funded by the CHIPS and Science Act. Micron is one of the biggest beneficiaries of CHIPS Act funding, and the Building Chips in America Act passed by the US House of Representatives Monday would allow it to access funding for its projects in Idaho and New York faster.
It is quite transparent that these companies cannot make enough chips in the short term and tech companies are throwing money at them to try to produce the supply that is required for the AI build-out.
Whatever you think of how many AI chips will be needed to deploy AI in full capacity - the real number will dwarf that.
The energy generation needed to power this new technology is so immense that it could even raise the temperature of the earth a few degrees from the sheer energy it will emit.
We are at the beginning of the AI revolution and the chips are currently the best way to play it.
I am bullish chip companies who produces AI chips.
https://madhedgefundtrader.com/wp-content/uploads/2019/05/cropped-mad-hedge-logo-transparent-192x192_f9578834168ba24df3eb53916a12c882.png00april@madhedgefundtrader.comhttps://madhedgefundtrader.com/wp-content/uploads/2019/05/cropped-mad-hedge-logo-transparent-192x192_f9578834168ba24df3eb53916a12c882.pngapril@madhedgefundtrader.com2024-09-27 14:02:442024-09-27 14:52:18Chips Shine Through Again
Below please find subscribers’ Q&A for the September 25 Mad Hedge Fund Trader Global Strategy Webinar, broadcast from Lake Tahoe Nevada.
Q: The iShares 20+ Year Treasury Bond ETF (TLT) is not advancing like I had hoped. I’m not sure why the interest rate cuts have not impacted the 20-year maturity—is it too far out?
A: It’s not an issue of maturity; the fact is that the market has been discounting falling interest rates for six months, all the way back to March. It’s a classic “buy the rumor, sell the news” scenario. (TLT) rose $20 off the low this year, and once the rate cut actually happened, all the news was in. That is why I actually went short the TLT a couple of days ago, and that trade immediately started making money. Here’s the real problem: Fed futures are discounting 250 basis points in rate cuts by June of next year. If you don’t think we’re going to get 250 basis points in rate cuts, which is two 50 basis point rate cuts and five 25 basis point rate cuts, then the market is overbought for the short term and we’re selling short. That’s exactly what I did.
Q: Is it too late to buy Tesla (TSLA) and Nvidia (NVDA)?
A: No, it’s not, I think Tesla could hit $300 this year, and Nvidia could revisit $140. However, the more you wait, the more pain you have to take along the way. Nvidia did drop 40% off its high at one point this year, and Tesla dropped 80% off its high. The price of coming in late is pain, so be ready to take that pain or, even worse, to stop out.
Q: What is your take on Japan’s attempt to take over US Steel (X)?
A: Well, it’s entirely political. They definitely picked the wrong year to take a run at US steel because it’s headquartered in Pittsburgh, Pennsylvania, and neither political party can win their election without winning Pennsylvania. Nippon Steel is now 3x larger than US Steel (I covered the company for ten years when I lived in Japan.) It’s the steel factor Jimmy Doolittle bombed in the Pearl Harbor movie. US Steel is using 140-year-old technology—Open Hearth Technology—which hasn’t been updated since the Great Depression. Nippon Steel, meanwhile, is promising to scrap all of that and bring the Steel Industry into the 21st Century. All great ideas for Nippon Steel and their shareholders, but not so great for Unions; all of these takeovers always result in massive layoffs of Union workers. So, that is the issue. That’s where a large part of the added value comes from.
Q: What are the chances that interest rates drop to zero?
A: Zero. I don’t think we’ll ever see 0% interest rates again because people now understand the massive damage that causes to the economy and to savers. So, on the next interest rate cycle, we’ll go down maybe to 2% if we get a recession, but probably not much more than that.
Q: Is it a good time to buy FedEx Corp (FDX)?
A: Yes, it probably is. If there was one rule of trading this year, you buy everything on top of these monster selloffs that are caused by weak guidance. We did it on Palo Alto Networks (PANW) earlier this year—people made a fortune on that. FedEx just did the same thing, so yes, I’m looking very carefully at FedEx calls, call spreads, and LEAPS two years out.
Q: I recently saw a recommendation to buy California Utility Company PG&E (PGE) because of recent revenue gains. Should I take a look?
A: Absolutely, you should. PG&E has gone bankrupt twice in the last 25 years, and the current new management seems to know what they’re doing. They borrowed $20 billion to underground all the long-distance power lines in the state so they won’t be liable for any of these gigantic wildfires that caused the last bankruptcy. Also, you kind of want to own utilities when interest rates are falling because utilities are among the biggest borrowers in the country.
Q: Is Global X Uranium ETF (URA) a good proxy for Cameco Corp (CCJ)?
A: Yes, another one is Consolidation Energy Corp. (CEG), but they’ve all had absolutely astronomical moves ever since the announcement came out that Microsoft was reopening the Three Mile Island nuclear power plant. So, wait for a dip, but the thing is just going up every day right now.
Q: Is it time to buy iShares 20+ Year Treasury Bond ETF (TLT) LEAPS?
A: No, LEAPS territory was last year or the beginning of this year when we were in the $80s (and we issued a ton of (TLT) LEAPS last year.) LEAPS are what you do at market bottoms, not at new all-time highs or two-year highs. Remember, if LEAPS don’t work, they can go to zero, and you want to avoid the zero outcome as much as possible.
Q: Should I look at Visa Inc (V)?
A: Yes, this is another one of those poor guidance situations leading to 20% selloffs. In Visa’s case, they’re being sued by the US government for antitrust because they own 47% of the credit card market. So, I would maybe wait a little bit more, let the market fully digest that, and then Visa’s probably a really strong buy because they’re still growing at 15% a year and minting money like crazy.
Q: Do you see gold going to $3,000 next year?
A: Absolutely, yes, unless it goes to $3,000 this year, which raises a better question: what happens when gold hits $3,000? It goes to 4$,500, because Chinese savers have no other place to put their money except gold. The real estate has crashed and isn’t coming back, they don’t trust their own banks or currency—there really is nowhere else for them to put their own money. They don’t even buy gold miners, they just buy the gold metal and coins. So I think we could see much higher highs than gold, and I’m sticking to my longs.
Q: Will silver continue to lag?
A: No. In fact, in the last couple of weeks, silver has done a big catch-up that is happening because recession fears are going away. Even the soft-landing fears are starting to vaporize—we may have no landing at all. The economy may just keep going, and silver is far more sensitive to the economy than gold is; and that is all silver positive. When we get to the metals, you’ll see how much silver has actually caught up. Silver is probably the better buy here because it tends to outperform gold by two to one.
Q: Do you think the Japanese will cross 100 yen to the dollar in the near future?
A: No, but I think it may cross 100 to the dollar in two years. You’re looking at a permanently weak US dollar from now on. As long as we’re cutting interest rates faster than anyone else, our currency will be the weakest. Japan’s rates are at zero, so they’re not going to cut interest rates at all, which is why we've had this enormous move in the Japanese yen.
Q: Can you give me some good renewable energy stocks and reasons why they are good buys?
A: Well, my favorite renewables are the Canadian Uranium stock Cameco Corporation (CCJ), First Solar (FSLR), which has been the leading industrial-scale solar producer for a long time, and NextEra Energy (NEE), which is very heavily dependent on producing electric power from renewables and also have a 3% dividend.
Q: Why is the euro going up even though their economy is in such terrible shape?
A: Europe has much lower interest rates than the US, and therefore, much less ability to cut interest rates than the US; it is the interest rate cuts that are driving currencies down, and we are the world’s greatest interest rates cutter right now. So, that is why you’re getting outperformance of the euro (FXE).
Q: Financials have moved up over the last two weeks; what’s your take on year-end and beyond? Should I buy Goldman Sachs (GS), JP Morgan (JPM) and Morgan Stanley (MS)?
A: Yes on all three. They’re all big beneficiaries of falling interest rates, improving economies, declining default rates, and rising stock markets. So, you have a triple play on all three of those. I’d be buying the dips on all financials.
Q: When will the sell volatility come back?
A: When you get the Volatility Index ($VIX) over $30. That seems to be the sweet spot for selling volatility. We are now at $15.
Q: If the US sharply increases tariffs, what will be the impact on the economy?
A: It would basically amount to a 20% price increase on everything you buy—from clothes to electronic parts to everything else—and the stock market would crash. Probably 90% of the non-food items Walmart (WMT) sells is from China. That’s why they call it the Chinese embassy. Tariffs are a tremendous restraint of trade and never, ever work, except for targeted items like cars or solar panels. For instance, I am in favor of a 100% tariff on Chinese cars to keep them from demolishing our own car industry as they are currently doing in Europe.
Q: Do we expect commodities like copper (FCX) and foodstuffs to go up as rates are cut?
A: I do. They’re big beneficiaries of falling rates, but more importantly, they’re even bigger beneficiaries of a stimulated Chinese economy, and that’s why we see these monster moves over the last two days.
Q: If you had to invest in one rideshare company, would it be Lyft (LYFT) or Uber (UBER)?
A: Uber—they have far superior management, they’ll be the first into robo-taxis, and they are constantly evolving their model, with Lyft always struggling to catch up.
Q: How will antitrust regulation affect the Magnificent Seven?
A: The bottom line is it will double the value of the Magnificent Seven. If these companies are broken up, the individual parts are worth far more than the whole companies, and we saw this when we broke up AT&T (T) 50 years ago, and the resulting seven companies within a year had a combined market value that vastly exceeded the original AT&T. I actually participated in that deal when I was at Morgan Stanley (since I am 6’4” I was asked to carry the ballots from one floor to another). Expect the same to happen with the Magnificent Seven. They will be worth double or triple more.
Q: If China has a falling population, how will a stimulus program help?
A: Well, it will fill in for the 600 million consumers who were never born as a result of the one-child policy. Not many others are talking about this besides me, but the fact is that the current economic weakness comes entirely from the one-child policy, and there is no way out of that, so they are going to have to keep stimulating again and again, much like the US did through the pandemic.
Q: If you can buy gold and silver on the UK market in sterling, does that make more sense for a UK resident?
A: Yes, it does, since your home currency is in sterling. You will actually get a double play or a “hockey stick effect” because not only is gold going up against the US dollar, but sterling (FXB) is going up against the US dollar, so you’ll get a multiplied effect relative to the pound. We used to play this all day long in Europe in the 1970s and 1980s, back when you had individual currencies to trade and the euro hadn’t been invented yet.
To watch a replay of this webinar with all the charts, bells, whistles, and classic rock music, just log in to www.madhedgefundtrader.com , go to MY ACCOUNT, click on GLOBAL TRADING DISPATCH, then WEBINARS, and all the webinars from the last 12 years are there in all their glory.
Good Luck and Good Trading
John Thomas
CEO & Publisher
The Diary of a Mad Hedge Fund Trader
Oracle plans to increase their amount of AI data centers from its current 85 to 2,000.
That is the most important number to take away from an analysts meeting with Oracle management.
Readers should ride on the coattails of this AI data center firm as throw billions upon billion at increasing the amount of AI infrastructure.
Readers absolutely need to know that a great swath of tech is dead and not innovating - growth rates collapsing faster than the U.S. birth rate.
It is important to position yourself at the cutting edge of innovation and growth and that is precisely companies who are knee deep in AI data center infrastructure investments that includes chip companies that produce GPUs like Nvidia.
In fact, Nvidia supplies Oracle and most other tech companies with data center chips called graphics processing units (GPU).
Nvidia has experienced an eye-popping surge in its revenue over the past year, and GPU demand continues to outstrip supply.
Oracle's data centers are unique because they are automated. Each one is operationally identical regardless of its size, and since they don't require human workers, it allows the company to build them quickly. Plus, Oracle's RDMA (random direct memory access) GPU networking technology allows data to flow from one point to another more quickly than traditional Ethernet networks.
Oracle has 85 data centers up and running with 77 more under construction as of the end of August.
Next year, Oracle intends to offer a cluster of 131,072 GPUs, which is a big step up from its largest clusters now, at around 32,000 GPUs. But there's another difference:
The new cluster will use Nvidia's latest Blackwell chips, which can perform AI inference at 30 times the pace of its flagship H100, which Oracle currently uses. Theoretically, it's going to allow developers to build the largest AI models in history.
In fact, Oracle spent $6.9 billion on data center infrastructure in 2024.
Oracle is going after the best technology in Nvidia’s Blackwell chip which is a solid reason to get interested in Oracle stock.
I don’t believe AI infrastructure spend will dissipate anytime soon and as the rest of the tech sub-sector growth falters, this one little area of AI will hold up the rest of tech.
This is why we are seeing extreme concentration of outperformance in just a handful of tech names and I don’t believe we will experience a scenario of spreading the wealth around to the less growth oriented subsectors.
In fact, I think the concentration will become even more outsized in a handful of names as a winner takes all mentality wins out in the tech sector.
We are just scratching the surface in what will become a massive explosion of AI data centers everywhere to satisfy the extreme demand of computing that it will require to pull this off.
Nothing indicates that this would be the wrong trend to follow and that assumption follows through to the astronomically high stock prices of the companies involved.
Oracle is one of these companies that readers should not dismiss.
It is at the heart of the AI infrastructure story that has legs.
https://madhedgefundtrader.com/wp-content/uploads/2019/05/cropped-mad-hedge-logo-transparent-192x192_f9578834168ba24df3eb53916a12c882.png00april@madhedgefundtrader.comhttps://madhedgefundtrader.com/wp-content/uploads/2019/05/cropped-mad-hedge-logo-transparent-192x192_f9578834168ba24df3eb53916a12c882.pngapril@madhedgefundtrader.com2024-09-25 14:02:422024-09-25 15:16:03From 85 to 2,000 AI Data Centers
Global Market Comments
September 23, 2024 Fiat Lux
Featured Trade:
Featured Trade:
(MARKET OUTLOOK FOR THE WEEK AHEAD or THE DOCTOR JEKYLL AND MR. HYDE MARKET),
(NVDA), (MSFT), (GLD), (NEM), (TSLA). (CCJ), (DHI), (TLT)
There is a very high degree of risk involved in trading. Past results are not indicative of future returns. MadHedgeFundTrader.com and all individuals affiliated with this site assume no responsibilities for your trading and investment results. The indicators, strategies, columns, articles and all other features are for educational purposes only and should not be construed as investment advice. Information for futures trading observations are obtained from sources believed to be reliable, but we do not warrant its completeness or accuracy, or warrant any results from the use of the information. Your use of the trading observations is entirely at your own risk and it is your sole responsibility to evaluate the accuracy, completeness and usefulness of the information. You must assess the risk of any trade with your broker and make your own independent decisions regarding any securities mentioned herein. Affiliates of MadHedgeFundTrader.com may have a position or effect transactions in the securities described herein (or options thereon) and/or otherwise employ trading strategies that may be consistent or inconsistent with the provided strategies.
