@@TickerSymbolYOU Someone just called you out to gain views on Supermicro.... I tried and defend you as trading is a fluid game... Name of the channel is Chris Norlund .... Take a look ... and let me know your thoughts....
Your predictions use a linear trend. They are closer to a hyperbolic. Slow growth large growth slow growth. AMD is in the initial slow growth period. Adoption, like Epyc will eventually exponentially change upward.
I don't think the argument for AMD is that they're going to overtake NVDA, at least not anytime soon, but rather that they can build a highly profitable business and carve out meaningful market share as the #2 GPU provider.
It's the CPU's that are the real winners with AMD. their GPU's are more like a multi role fighter plane and the NVIDIA ones more like interceptors. Sure Nvidia has a monster GPU's but they're pricing them out of the hand of most people. AMD is allot cheaper. dollar for dollar, AMD every time.
Nvidia with it's 600W GPU is nice, but nah I'm sticking with AMD on my next build. As long as I can get 4k ultra 60-120hz, I'm good. Plus the fact that they have a monster CPU and FSR makes any GPU out there last for a while.
It’s hard to believe NVDA can get bigger with such an astronomical market cap, but the public hasn’t really even seen AI yet or AI hardware/real estate demand
The money is not in the customer front mom and pop buying a GPU every 5 years for Timmy at home. It's in B2B server space, where major cloud providers are constantly out of GPUs in every major region.
@@SniperMayer the B100/200 infrastucture is a generation ahead because they were building up the infrastucture and software around AI a while back - AMD is stuck playing catch up, til eventually Nvidia is going to be further and further ahead on the bell curve.
AI is an overhyped nothing burger. We are approaching the phase where most of these overhyped promises will crash and burn because they are technically not feasible.
Respectfully, the entire town of Ashburn, Virginia is like 80% data centers full of AI gpus. The real estate demands are continuing to grow, but people have seen it in Northern Virginia. Regarding hardware demands, there are global backorders on commercial gpus
Let's think AI is like a "hole in the ground" business. Today everyone is trying to buy the best shovel to dig a hole. In time companies will realize that it doesn't matter how fast you dig a hole, it's more important the location of the hole and how wide and deep it is, then they will realize, many shovels will not help digging a better hole and then the shovel business will go for the best shovel per dollar and not the fastest shovel. This is my analogy with AI, today everyone is buying GPUs (shovels) without a proper plan to monetize the AI service (dig a good hole) and the race is letting everyone blind to the fact that companies need a good AI and not the first AI. When they realize this, they will no longer buy the more expensive GPU and focus on data, and the execution of a lucrative AI and for that the better tool is a cost effective tool. It doesn't matter who has the most market share today when the race is between headless chicken, what it matters is, who will in the future, sell a great platform per dollar spent, to execute the commodity that AI will be... Besides NVidia is building chips around the limit area (mm²) of today's technology. If they don't find a way to pivot from bigger and more power hungry strategy, they will be limited in future iterations of their GPUs that can't grow in size and need to evolve just with architecture improvements where, today, AMD has room to grow in chip size and architecture improvements. Time will tell... Just my 2 cents about the tumbnail saying "it's over"
Great metaphor and thinking. Time will tell. You could also add that man of Nvidia client are not very happy with the quasi monopoly of Nvidia and that they favor more competition.
I can get on board with this analogy. Especially since 3 big AI Tech company CEOs have just said the data ingestion of AI has capped out and all the low-hanging fruit has been had. Which to me (and my minor understanding) tells me that the speed at which the data is trained doesn't inherently mean you have a better or more advanced product. AMD still has a dog in this fight because their chips suck a lot less power and cost quite a bit less as well. Which to me means smaller more value-minded start-ups could upset some of these larger companies if they can get a foothold and create a superior product with the AI powered by cheaper less power hungry AMD units. Then again I could just be a moron hahaha .🤷
AMD has a huge upside at some point. Nvidia will always be AI king but if AMD can offer a value option for AI customers their value could 5x very quickly. especially with its stock price falling. Local AI chips will be important as well. AMD will have a place to play in this AI revolution.
The problem is their value is not better due to way worse software & support. AI engineers are expensive, and you'd need more to do the same on AMD vs Nvidia.
I have to disagree when you say AMD/Intel doubles down on older technology instead of building ARM. Yes: Today's big cloud company are interested in ARM architecture, but they alone cannot cover all uses cases. Under the hood, today's x86 CPU already share many of architecture design choice has ARM : simple instruction set, pipelining, out-of-order execution (AMD), superscaler capability. Via micro-code, what intel and AMD built is essentially a RISC computer that emulates CISC. They already have their patent that give similar result to ARM architecture, I don't know why you would them to "pivot" to ARM... Moreover, AMD/Intel's SIMD instructions have wider and more features than current ARM processors. Many software and upcoming projects will rely on extended SIMD feature which is not available on ARM. For example: newer blockchain technology will use SIMD instructions extensively on x86 CPU not ARM. Also, while ARM is supposed to be RISC (R for "reduced"), modern ARM do have a shitload of instructions, they are getting themselves more complex. When ARM started in the 80s: they has 50 instructions, now ARMv8-9 has up to 1500 instructions. Which is getting closer to x86 2000/3000 instructions. Finally, Intel did invest in "RISK-V" which share part of the original ethos of ARM: reduced instructions set but more modular.
Correct. Strengthening x86-platform has more to do with protecting patents and brand than it has to do with AMD being unable to make an ARM-compatible CPU. They absolutely can and likely have already done at least on experimental one but keeping it internally for evaluation. To most people CPUs are simply magic boxes that does magic, but internally they are all quite similar regardless of architecture because magic can only be done in so many ways unless you want your CPU to burn.
@@torgnyandersson403 You are correct. AMD did look into ARM as a possible product when Jim Keller was there. That was canceled at some point, but unofficially, there are rumors that AMD works currently on an APU that uses ARM CPU. They did bid against Nvidia for Switch 2(they lost the bid tho), so since Switch uses ARM it makes sense the rumors are likely true
Great video-very informative. I've been in nvda since 2016. I'm regularly watching a number of other tech stock analysts, and you have brought up nvda vs. amd issues I haven't heard from anyone else--good work! You're now on my 'regular' list.
I don't think ARM chips will ever replace x86 as a whole. Not after ARM suing their biggest customer Qualcom. That move probably scared away a lot of potential customers right into the arms (no pun intended) of RISC5 which has a free license or to stick with x86.
why people saying the gap between Nvdia and Amd are getting wider. amd instinct gpu sales are up 50% qoq from 1b in q2 to more than 1.5b in q3, and it will hit more than 2b in q4. i doubt Nvidia’s hopper/blackwell right now has that qoq growth
Just because NVDA has more of its revenue from AI and AMD has less today doesn’t mean it will stay that way. Your argument for dollars worth is only valid today, in the future amd revenue could come from a greater % of ai
@@lllcool2002 Exactly, and I wouldn’t be astonished if they push those earnings back either. However I agree, I think they could blow all earning numbers out of the water and because of the situation, investors won’t care because their integrity is in question.
It seems to me that AMD will do well. Firstly, the demand for AI chips will be greater than NVIDIA will be able to handle, and AMD is the only alternative. Secondly, AMD is introducing a rather significant change that is not widely known. NVIDIA's success was based on the fact that it has the same CUDA cores for consumers and data centers. Thanks to this, programmers were more willing to optimize software for NVIDIA cards. AMD has only now realized its mistake and is introducing a single core standard for consumer and server cards, while at the same time establishing cooperation with AI startups. They may never surpass NVIDIA, but technologically they are still in the game. They mainly need to work on popularizing their technology among ordinary users and not just large corporations. When people discover that stationary applications such as Stable Difusion or various LLMs work just as well on AMD cards, only for less money, programmers will be more willing to optimize models for their cards and the fight will be even.
AI stocks are driving the future, and NVIDIA's dominance in GPUs keeps them ahead. AMD has potential, but NVIDIA's supply issues resolving could widen the gap. I'm watching both closely for 2025 opportunities.
I agree. Even with great opportunities, we should proceed cautiously. Seeking market analysis or advice from certified market strategists is important.
People betting on NVDA right now to become a $10tn market cap company are going to get taken to the cleaners when they realize the AI CAPEX cycle won’t grow to infinity and is instead cyclical.
I've got a few things to clear up and share some key details with you. My number breaks down like this: First, plus 1215... Next, 403... And finally, 4030. I'll be waiting for your W.A message.
AMD are a very well run company. However I think trying to catch NVD in AI is like trying to create a social media platform that competes with Facebook. NVD is taking the AI market and solidifying its position. If NVD have a mis step then AMD could catch.
We could be seeing that mistep soon. Keep in mind back before the AI hype bubble Nvdia has had a long track record of lying cheating and scamming. I don't think most people quite understand how deep SMCI and NVDAs ties go, not only do they have very close business ties but Huang and Liang are literal best friends. . I'm beginning to suspect why SMCI can't release numbers and their accountant just quit is because they have no real way of hiding the fact they were helping NVDA spoof their obviously fraudulent numbers. Lets not pretend NVDAs other customers aren't in on it as well. Wouldn't be the first time NVDA has put out fraudulent numbers.
And these prices buying NVIDIA stock is the worst idea. Especially taking into consideration all the latest hints there is a problem with simple scaling. 🤯
I wanted to stop and thank-you for all your videos! You have helped me find some true gems. I have owned NVDA for 20 years (thus the intrust to start). But I never would have found PLTR in time for this years push. Again Thanks and glad to see you back up and running!
In conclusion, the Tickersymbolyou suggests holding AMD if already invested, but for investors with a strong focus on AI, Nvidia offers a more compelling growth story due to its market dominance and rapid innovation cycle.
Perhaps you should research the AMD Versal solution a bit more. It is an FPGA, ARM core & GPU merged into a single platform. FPGA's are superior in many ways compared to GPU's. Low-Latency Customization: FPGAs, like those within the Versal, allow for custom data paths and parallelized logic that can be tailored precisely for specific low-latency tasks. Since the hardware can be configured to execute specific functions without needing to pass through layers of software, FPGAs typically provide lower latency than GPUs, which rely on general-purpose cores. Deterministic Performance: The Versal platform, with its FPGA component, allows for highly predictable, deterministic performance, which is essential for latency-sensitive applications (e.g., real-time signal processing, high-frequency trading, or edge AI). Deterministic response times are often harder to achieve with GPUs, as their performance can vary based on workload and memory access patterns. Real-Time Data Processing: For tasks requiring immediate, responsive processing, such as those in telecommunications or automotive applications, the Versal’s mix of FPGA, ARM cores, and AI engines can handle data streams and execute logic in real-time. In comparison, GPUs, despite their high throughput, typically introduce more latency due to their software stack and reliance on batching operations to maximize efficiency. Direct Hardware Control: With the Versal, developers can set up data flows directly within the hardware, reducing the need for time-consuming data transfers or memory bottlenecks. This control over the hardware can be critical for latency-sensitive applications, allowing FPGAs to outperform GPUs in these scenarios. Power Efficiency in Low-Latency Workloads: In latency-sensitive scenarios, the Versal solution can perform specific tasks more efficiently in terms of both power and speed, as it avoids the general-purpose, high-power approach of a GPU. Blackwell GPUs are optimized for sustained throughput over latency and often run best with a certain workload volume that compensates for their higher power requirements. NVDA blackwell should be considered a general workhorse with massive parallelization abilities at the cost of latency and customization. AMD Versal is superior for latency sensitive applications, GPU's will never dominate this space. AMD Versal will eat market share of Blackwell and for a lot of real time sensitive or edge products, will be the only option considered as Blackwell is weak in this area.
@@mcspeed07able seems like a solid justification , however at times stock movement is like fashion industry that follows trend (or) like movies with famous actors make box office than movies with good story content ... therefore from stock point of view if NVDIA falls rest will fall - if NVDIA moves up rest may follow ...
@@mavensept except time and time again we see the exact opposite, with Nvidia falling and AMD taking prominence. Yet the vice versa happens and we act like AMD is dead all of a sudden lol
@@mcspeed07able Doesn't matter what AMD has. The world has moved on. VHS sucked, but once it took hold of norms it didn't matter if Betamax was superior.
@@RajDeelish You are extremely wrong to use that example in this case. The Mega Cap tech companies will quickly move to whatever option is best and are actively working on custom chips with various partners. This industry can do a sharp turn at a moments notice so don't get complacent.
Having programable ASIC style chips from xilinx is a major game changer for AMD it was definitely a positive acquisition for them and put INTC on notice. Nvidia Mellanox acquisition was a huge game changer for them as well.
Do you really think META, GOOG, TSLA, MSFT, OpenAI, AMZN, APPL are going to let NVDA be worth more than all of them combined? I think over time these companies will a) Create their own chip b) help AMD (especially in software area to catch up to CUDA) compete with NVDA.
apple and intel are trying to make their own chips, and amd is trying to catch up to where nvidia is RIGHT NOW - while nvidia is already building the next big thing, nvidia gambled 10 years ago on AI by building up their software and the entire architecture around AI - not just the chip but all the parts the chip relies on, the b100/200 isn't just gpu's - it was rebuilt from the ground up & inside out to handle the better gpu's.
This is all true uintill the AI bubble holds. By now having a 150000 GPU cluster that powers an LLM that can barely give you some resemblance of thinking is pretty underwelming.
AMD also won’t tank from an AI bubble burst if one happens, especially compared to how much NVIDIA would due to it being the lions share of their company
Datacenters are a sleeper market that people forget about. We forget that the internet lives on physical computers, thats what datacenters are. They are already huge and will be part of out infrastructure indefinitely. Targeting those in whatever way seems valuable will be a good investment imo
You channel is like comfort food. I tend to get your recomended videos after I make a switch in stocks. I got rid of shares of apple since I don't think they are innovating any more. I found Nvidia, AMD, FLEX, and Tesla, at the start of the year being a good choice. I already had google, amazon, microsoft, so these were a good swap for apple. This year was great, but thanks making me keep a closer eye on AMD.
Hi Alex! Please make a video comparing Vistra, Constellation and Talen Energy companies. I'm sure your viewers would love that! Powering AI, and EV cars of the future. The stocks have already taken off. What's your take on all of this! Especially on Talen! Too late to get in?
Also I think for PE ratios it is 47 to 42 if I am not mistaken which means AMD stock is almost as highly valued as NVDA. I would be more into AMD stock if its PE was in the mid to low 30s especially when I couple that with all the other points you made on the video.
Look at forward P/E, not trailing P/E... AMD forward p/e 28.49 NVDA forward p/e 34.60... PEG ratio AMD (5yr .38) NVDA (5yr 1.05). We saw what happened to NVDA when they started realizing profits from data center sales... AMD is at the beginning of that same ramp now...
@@mr2d2@mr2d2 Good point, and honestly, I do not know as much about comparing forward P/E ratios but even with that said your point about the AMD ramp up is a good one to think about = POTENTIAL is significant for sure! I still believe NVDA has 4 quarters of uninterrupted ramp-up in all aspects of sales and the stocks, but I will keep some money in AMD stock also. I guess we will find out on November 20th if NVDA is still hitting it out of the park.
@@mr2d2 you said it’s the same as NVidia when they started realising profits from data center sales. it’s not the same. NVidia had no competition. AMD has to compete with a superior competitor for these sales.
Sounds like you lost a lot on SMCI stock. I did as well. About half of what I've invested into the company before I got out. Things were going great until we learned that SMCI cooked the books. No one knew what SMCI did until it was too late. I don't blame TS You. Motley Fool, in there Hidden Gems publications pushed five Chinese stocks. Three plummeted on the same day when it was learned that all five companies had cooked the books. Again, I lost a butt ton of money. I didn't Blame Motley Fool because they lost as much as I did. Even if I invest into Nvidia, the stock would have to more than double for me to reclaim my losses but that's exactly what I am going to do. Wish you well in your endeavor.
Arm is not necessarily the architecture of tomorrow. They are alternatives that are still in development to x86, because no one could product x86 due to licensincing. X86 has really all the software support, which arm is catching up, tho its a long long eay to go. Performance of x 86 per mm and connectivity on epyc is still unmatched
@Ticker Symbol: YOU why does your video feel so "lip-synched"? Do you extra voice over for your video? Or is it a slight delay between video and audio? Feels so uncanny.
@@TickerSymbolYOU Now that I read this comment, I see it too, but I'm not sure if it's a psychological effect. If there is a delay it could be as little as a frame or two.
Nvidia has the software stack (Cuda) as the standard. Amd doesn't have access to that API (without a huge amount of effort (ROCM)), talent pool and technology. So their GPU is not competitive. Though their CPUs are very competitive in the server market.
I initially did the big ticket buy, 125k into SCHD, 75k TSLA, 25k VYM, 25K VUG. Now I'm dca buying roughly 2k every week of whatever is on sale, and looking to add more tech positions to my portfolio. I'm looking to hold long term 15 - 20 years, so hopefully my lump sum buy in doesn't bite me in the ass long term.
In the past month, my "unexciting" index funds provided me with over $6,000 in dividends, giving me the option to spend without selling shares. Currently, I've opted to reinvest the dividends to acquire additional index funds for future growth.
@JG-nm9zk Just wait until the regular Wallstreet humans start talking about the software growth in revenue ... maybe you remember this ....I knew that GPU would be the brain of Ai computing 10 years before anyone was talking about it.....people were all saying it's a gaming business..lol
It all comes down to cost. That's the bottom line for any business. How much does Bank of America want to spend on a chat-bot to answer customer questions? Less then the call center employees. Driver-less cars? They still can't get that to work. (Even on trains).
The interconnect issue you mention is crucially important, it's not glamorous to talk about it, but it IS how we scale data centers. Ethernet is simply superior in cost of implementation, universality and labor force specialization and support. AMD has actively collaborated with the group and is well positioned to leverage that scale and this leaves the entire industry wide open for their solutions. Nvidia preferred the thunderbolt approach. We all know how that ended up. AMD and their approach means they WILL capture a larger share of the market. Its just simply going to happen.
For general computing, RISC V may be the winner in the future. ARM/Qualcomm IP issue got others looking another way. So will see how long ARM will run servers, Single Board Computers, Smartphones, etc. As for AI. NPU / AI ASICs on the same process node X nm will have better performance per watt compared to GPUs. So I could place a bet on Cerebras. I want to see more ASICs for general-purpose compression like ZSTD, Brotli, and others. Cast Inc. is a player in compression/decompression accelerators.
I hear what you’re saying. I would have stayed invested in AMD but I have so many more lucrative stocks to put the $ in…as well as I’m not convicted that Intel was a good decision. Thanks, Alex!
There's no reason to make ARM chips in 2024. Instruction sets make up less than 1% of modern CPU's. All hardware designers are diversifying their compute units to offer more specialisation (i.e. larger instruction sets along RISC principles). Energy efficiency of ARM chips can be fully explaioned through them usually being one processing node ahead and the focus on SoC integration, which is something AMD has a lot of experience with.
I always keep hearing from Super Investors, it's always far better to be invested into the #1 clear leader. And not to chase after the #2 or #3. Who is the clear leader in the AI sector? We know.
Yeah, and AMD will never be able to even come close to touching Intel, especially in the server room where the biggest profits are - that's what talking heads were saying just a few years ago. AMD has already announced (over a year ago) that they are developing ARM chips, and I would wager that their outside the box thinking will come up with market leading ARM chips in short order.
People forget desktop market is not a laptop market and it's not a corporate market. If you want to invest you should care mostly about corporate market (AI, servers, networking, datacenters, FPGAs etc). If you care about parts and their performance, look at hardware individually per platform and price range. I wrote this because I am both a tech investor and I love new consumer tech too, just about tired of people fanboying over specific corporations and behaving like cavemen.
Good presentation. However I think AMD AI revenue will grow as a percentage of sales going forward which will increase the overall growth rate of the company. Beside that the stock is selling as approximately 23 times non gap eps of 6 for 2025. Furthermore their gross margins will trend upward too. The issue with AMD is that people keeps comparing it with Nvidia the undisputed king of AI. It is a bit unfair. AMD was able to beat Intel with much less financial resources. It will be a different ball game with Nvida which as a very different and aggressive company. I think AMD could aim at 10-15% of the AI accelerator market within 2 years. Finally we can see that the accelerate product release of Nvidia has created some technical over heating issues that could help AMD. TIME WILL TELL. thks for the video again. Good work😊
Thank you so much for bring us such a high quality information. I have a question...are these superclusters still needed once the model has been trained? what happens to the hardware once training has been completed?
Training is never really "completed" -- models go through fine tuning and re-training as new data comes in while new models are constantly being built (either better/bigger general models or more application-specific ones)
IT speacialists are recommending AMD sell their old cards as cheap as posible and prepare lunch RX 8800 series as low as $400 if AMD want to keep at least 25% of market share from nVidia-nPolosi.
How do you feel about Taiwan Semiconductor? It seems more poised to benefit from AI spending than both Nvidia and AMD long term because companies like Google, Apple, and Meta can and likely will design their own chips
I'm keen on AMD's upcoming APUs it'll be so cool...once my current desktop dies I'll be moving onto a MINI PC setup. building desktop rigs is becoming expensive each time
Years ago when AMD was on the edge of bankruptcy, I bought about 25,000 shares at $1.50, but I was chickened out and sold after it dip to bottom and went back up. However, I bought NVDA after that. In 2022 I bought some AMD at around $65. In late 2023 when Lisa gave guidance 2B when Nvidia could not made enough chips to sell, I said to myself this is not good for AMD ecosystem for the years to come, so I sold all AMD (sold TSLA, too bc Elon promised and not delivered) and added more NVDA shares. That amount of money has now been 3x while TSLA and AMD are going nowhere.
Nvidia is buying back shares. AMD is diluting shareholders with stock based compensation. Nvidia has stock based compensation as well but they are buying back more.
Why do you assume they are wanting to catch-up - horses for courses. You don´t have to be the biggest - to be the best. This is a US mindset - but the rest of the World we go for quality rather than pack em cheap - sell them high mentality.
Ok. How about "harder to compete?" When it comes to accelerated computing and AI, harder to compete also means demise because AI progresses a hundred times if not thousands time faster than technology in the past. It's not like personal computing and games where you just bring some little chips to the table. It needs the whole stacks, platform, ecosystem, and partnerships. I have not even gone into arguing about when physical AI comes, AMD will have nothing to compete with Nvidia's Omniverse. I have subscribed to 2-minute-papers TH-cam channel for almost 10 years. Never heard of any AI paper from AMD while papers from Nvidia presented almost every week. All the billionaires and giant companies that go after Nvidia's chips and technology send me a message. Don't underestimate billionaires bc they are billionaires for a reason.
@@notbaltic280 The so called investors in the US only look at quarter to quarter. Most are just traders not real investors. We have been conditioned by the talking heads and the so call analysts that are by and large often wrong. It seems buy and hold is dead, which really hurts any real investor at least short term. I talked to a investor relation person one time and he told me they prefer institutional investors over retail. Retail is always the most irrational, I assume because institutional is not using their own money.
I'm concerned for the long-term future of x86. Apple's custom ARM chips are very good and it's only a matter of time until there's a huge disruption to the Windows PC market. The Mediatek+Nvidia partnership is interesting but they may not be working on high-end PC SOC's, just mobile/laptop SOC's. It'll likely be better than Qualcomm's recent attempt at ARM on Windows.
NVIDiA covers all aspects of the AI revolution. You can’t become involved in AI as a company and not include NVIDiA in the process! They are the dominant company in leading the AI revolution! It is a big part of my portfolio and will be the reason why I retire early! Good luck to everyone invested in NVIDiA!
Without CUDA alternative and stronger drivers development AMD has little chance against Nvidia in GPU and AI market. And this is not a problem since companies are closely connected.
It would’ve been also helpful if you included your thoughts on AMD’s acquisition of ZT Systems. Personally, this is a great deal that helps AMD gain more partnerships with hyperscalers, and AI infrastructure. Any thoughts?
Is AMD supply-chained to TSMC, feeding off whatever production capacity Nvidia has not consumed? Perhaps AMD should strengthen that strategic partnership with Intel to include foundry contracts.
I would suggest that AMD never had a headwind in supply, since both NVIDIA and AMD are reliant on TSMC, and to smaller suppliers; but there's 2 very important points that you're missing in your analysis. BOTH AMD and NVIDIA will be moving to a yearly release cadence. Also, because no one but AMD and Intel have X86 licenses (I believe VIA is the 3rd), they have to release ARM. But, it would be a mistake to think that neither company could release ARM processors themselves. They in fact, do; just not bread and butter chips
Most of my portfolio is in both of these and I'm winning huge!!!! I mean HUGE! Average yearly increase is about 120% yearly over 10 yrs-that's as far as I can get an average.
I have been investing for years. Unfortunately I missed the opportunity to jump off at $200, I should have used the AI hype. They simply cannot deliver like NV. That has become clear over the years. No question, AMD is a good company, they deliver well. Intel has done well, but unfortunately the profit is only made in AI and not PC or data center CPUs. The lag in software quality and support does the rest. Everything is hopeful, but it takes years to prevail. NV has simply beaten them so clearly, year after year.
Don't buy a single stock, just buy the entire sector. Because you cannot verifiably predict which stock will win the AI race, but for sure of of them will. Even people working directly in this industry cannot predict what will happen! When I worked with Nvidia hardware for AI research, I was sure it would never make it to the top of the AI hardware game, just because it's so full of glitches and errors. Yet, look where we are.
Probably not much. I imagine Nvidia has partners/installers willing to pick up the slack and then some. I'd imagine a minor slowdown in short-term sales and that's it.
One day those billions of investment in AI has to start making a "return of investment". I don't see where are they going to get that return from. Unless this is more of a "infastructure investment" in a conventional sense. "Skynet" is just over the horizon and it might be here sooner than we thought.
You can protect your privacy and support the channel by getting 20% off DeleteMe at joindeleteme.com/SYMBOL20
Alex, have you done a deep dive on AI on the mobile ecosystem ? Like Arm, Qualcom, Iphone ?
@@jmj48 yes I have! If you search for QCOM or Apple, you'll see my coverage of the Snapdragon X Elite vs Apple's silicon
@@TickerSymbolYOU Someone just called you out to gain views on Supermicro.... I tried and defend you as trading is a fluid game... Name of the channel is Chris Norlund
.... Take a look ... and let me know your thoughts....
Diversifying the range of production is important, 88% of AI production is dangerous.
Your predictions use a linear trend. They are closer to a hyperbolic. Slow growth large growth slow growth. AMD is in the initial slow growth period. Adoption, like Epyc will eventually exponentially change upward.
I don't think the argument for AMD is that they're going to overtake NVDA, at least not anytime soon, but rather that they can build a highly profitable business and carve out meaningful market share as the #2 GPU provider.
It's the CPU's that are the real winners with AMD. their GPU's are more like a multi role fighter plane and the NVIDIA ones more like interceptors. Sure Nvidia has a monster GPU's but they're pricing them out of the hand of most people. AMD is allot cheaper. dollar for dollar, AMD every time.
Remember yall, AI is big but the bread and butter of the world is server
@@TheMadMariner To be honest isn't this what they've done against Intel with the X86 and X86-64 architecture?
Nvidia with it's 600W GPU is nice, but nah I'm sticking with AMD on my next build. As long as I can get 4k ultra 60-120hz, I'm good. Plus the fact that they have a monster CPU and FSR makes any GPU out there last for a while.
@@lionelt.9124 Yes, exactly.
It's def not about taking over NVDA it's AMD gaining some share in Ai and building that out.
AMD gave us a blessing to load the boat i been praying for it to come down and Friday made my last big purchase
@@citizenm9590 I’ve just started nice to see your confidence I hope your getting a few more now :)
hows that working out lol
damn....it keeps going down. Is this the bottom?
I’m predicting around 120-115. Then buy. Or DCA into it over a week or so
It’s hard to believe NVDA can get bigger with such an astronomical market cap, but the public hasn’t really even seen AI yet or AI hardware/real estate demand
since they're now on DJI there are still people saying the stock has potential to double, which is mind blowing.
The money is not in the customer front mom and pop buying a GPU every 5 years for Timmy at home. It's in B2B server space, where major cloud providers are constantly out of GPUs in every major region.
@@SniperMayer the B100/200 infrastucture is a generation ahead because they were building up the infrastucture and software around AI a while back - AMD is stuck playing catch up, til eventually Nvidia is going to be further and further ahead on the bell curve.
AI is an overhyped nothing burger. We are approaching the phase where most of these overhyped promises will crash and burn because they are technically not feasible.
Respectfully, the entire town of Ashburn, Virginia is like 80% data centers full of AI gpus. The real estate demands are continuing to grow, but people have seen it in Northern Virginia. Regarding hardware demands, there are global backorders on commercial gpus
Let's think AI is like a "hole in the ground" business. Today everyone is trying to buy the best shovel to dig a hole. In time companies will realize that it doesn't matter how fast you dig a hole, it's more important the location of the hole and how wide and deep it is, then they will realize, many shovels will not help digging a better hole and then the shovel business will go for the best shovel per dollar and not the fastest shovel. This is my analogy with AI, today everyone is buying GPUs (shovels) without a proper plan to monetize the AI service (dig a good hole) and the race is letting everyone blind to the fact that companies need a good AI and not the first AI. When they realize this, they will no longer buy the more expensive GPU and focus on data, and the execution of a lucrative AI and for that the better tool is a cost effective tool. It doesn't matter who has the most market share today when the race is between headless chicken, what it matters is, who will in the future, sell a great platform per dollar spent, to execute the commodity that AI will be...
Besides NVidia is building chips around the limit area (mm²) of today's technology. If they don't find a way to pivot from bigger and more power hungry strategy, they will be limited in future iterations of their GPUs that can't grow in size and need to evolve just with architecture improvements where, today, AMD has room to grow in chip size and architecture improvements.
Time will tell... Just my 2 cents about the tumbnail saying "it's over"
Great metaphor and thinking. Time will tell. You could also add that man of Nvidia client are not very happy with the quasi monopoly of Nvidia and that they favor more competition.
I can get on board with this analogy. Especially since 3 big AI Tech company CEOs have just said the data ingestion of AI has capped out and all the low-hanging fruit has been had. Which to me (and my minor understanding) tells me that the speed at which the data is trained doesn't inherently mean you have a better or more advanced product. AMD still has a dog in this fight because their chips suck a lot less power and cost quite a bit less as well. Which to me means smaller more value-minded start-ups could upset some of these larger companies if they can get a foothold and create a superior product with the AI powered by cheaper less power hungry AMD units. Then again I could just be a moron hahaha .🤷
AMD has a huge upside at some point. Nvidia will always be AI king but if AMD can offer a value option for AI customers their value could 5x very quickly. especially with its stock price falling. Local AI chips will be important as well. AMD will have a place to play in this AI revolution.
The problem is their value is not better due to way worse software & support. AI engineers are expensive, and you'd need more to do the same on AMD vs Nvidia.
FUN FACT: Lisa of AMD and Jensen of Nvidia are first cousin
Yep, true.
Damn! didn't know that.Thanks for sharing
@@snaheelyadav5863 ye pass on the knowledge aye aye
Rumor! Lisa and Jensen both have denied that rumor.
Have her tell her cousin to buy AMD.
I have to disagree when you say AMD/Intel doubles down on older technology instead of building ARM.
Yes: Today's big cloud company are interested in ARM architecture, but they alone cannot cover all uses cases.
Under the hood, today's x86 CPU already share many of architecture design choice has ARM : simple instruction set, pipelining, out-of-order execution (AMD), superscaler capability.
Via micro-code, what intel and AMD built is essentially a RISC computer that emulates CISC.
They already have their patent that give similar result to ARM architecture, I don't know why you would them to "pivot" to ARM...
Moreover, AMD/Intel's SIMD instructions have wider and more features than current ARM processors.
Many software and upcoming projects will rely on extended SIMD feature which is not available on ARM.
For example: newer blockchain technology will use SIMD instructions extensively on x86 CPU not ARM.
Also, while ARM is supposed to be RISC (R for "reduced"), modern ARM do have a shitload of instructions, they are getting themselves more complex.
When ARM started in the 80s: they has 50 instructions, now ARMv8-9 has up to 1500 instructions. Which is getting closer to x86 2000/3000 instructions.
Finally, Intel did invest in "RISK-V" which share part of the original ethos of ARM: reduced instructions set but more modular.
Correct. Strengthening x86-platform has more to do with protecting patents and brand than it has to do with AMD being unable to make an ARM-compatible CPU. They absolutely can and likely have already done at least on experimental one but keeping it internally for evaluation.
To most people CPUs are simply magic boxes that does magic, but internally they are all quite similar regardless of architecture because magic can only be done in so many ways unless you want your CPU to burn.
Disagree in the sense that the results are miles away. ARM is ARM, let's not even compare
@ arm and x86 have similar implementation and similar performance. Yes they are comparable.
@@torgnyandersson403 You are correct. AMD did look into ARM as a possible product when Jim Keller was there. That was canceled at some point, but unofficially, there are rumors that AMD works currently on an APU that uses ARM CPU. They did bid against Nvidia for Switch 2(they lost the bid tho), so since Switch uses ARM it makes sense the rumors are likely true
Great video-very informative. I've been in nvda since 2016. I'm regularly watching a number of other tech stock analysts, and you have brought up nvda vs. amd issues I haven't heard from anyone else--good work! You're now on my 'regular' list.
Glad it was helpful! Thank you for taking the time.
aistockadvisor AI fixes this. AMD vs Nvidia stock debate.
I don't think ARM chips will ever replace x86 as a whole. Not after ARM suing their biggest customer Qualcom. That move probably scared away a lot of potential customers right into the arms (no pun intended) of RISC5 which has a free license or to stick with x86.
@@megapro125 riscv = reduced cost. Many component companies are committed to riscv because of the non complexity of licencing.
why people saying the gap between Nvdia and Amd are getting wider. amd instinct gpu sales are up 50% qoq from 1b in q2 to more than 1.5b in q3, and it will hit more than 2b in q4. i doubt Nvidia’s hopper/blackwell right now has that qoq growth
Why do you think companies like AMD or Intel etc want to catch up with Nvidia? You can carve out your own niche and still make a tidy profit.
@ i agree, i don’t think amd can catch up with Nvidia and i don’t think they need to
Just because NVDA has more of its revenue from AI and AMD has less today doesn’t mean it will stay that way. Your argument for dollars worth is only valid today, in the future amd revenue could come from a greater % of ai
Very true, but the trend is they're falling further behind (releasing new chips less often) AND starting on the back foot.
The AI sector Evolution just beginning. Nvda . Nvidia..AMD Dip . VHAI..19 % gains month.. Vocodia holdings..conversational AI.. Soun .9 % rise month.. SoundHound..Sym up .19 % month. Symbotic and more. Thumbs Up video/ comments. Thanks.
Any update on SMCI regarding their accounting firm quit on them?
I'm working on it right now
lol, the only news there is on that is out there already…. Not a lot of transparency on that yet…. Unfortunately.
@@TickerSymbolYOU Thanks, can't wait!
@@abidingdude222 well, earning coming out next Tuesday , but who is going to trust their numbers? Certainly not EY accounting firm, lol.
@@lllcool2002 Exactly, and I wouldn’t be astonished if they push those earnings back either. However I agree, I think they could blow all earning numbers out of the water and because of the situation, investors won’t care because their integrity is in question.
Have you in your archive of videos discussed which stocks related to nuclear energy and Ai may be of interest? Thanks
It seems to me that AMD will do well. Firstly, the demand for AI chips will be greater than NVIDIA will be able to handle, and AMD is the only alternative. Secondly, AMD is introducing a rather significant change that is not widely known. NVIDIA's success was based on the fact that it has the same CUDA cores for consumers and data centers. Thanks to this, programmers were more willing to optimize software for NVIDIA cards. AMD has only now realized its mistake and is introducing a single core standard for consumer and server cards, while at the same time establishing cooperation with AI startups. They may never surpass NVIDIA, but technologically they are still in the game. They mainly need to work on popularizing their technology among ordinary users and not just large corporations. When people discover that stationary applications such as Stable Difusion or various LLMs work just as well on AMD cards, only for less money, programmers will be more willing to optimize models for their cards and the fight will be even.
lisa su is a follower, not an innovator. May be she is just to way to conservative. Males account for the majority of day traders. Wonder why that is?
AI stocks are driving the future, and NVIDIA's dominance in GPUs keeps them ahead. AMD has potential, but NVIDIA's supply issues resolving could widen the gap. I'm watching both closely for 2025 opportunities.
I agree. Even with great opportunities, we should proceed cautiously. Seeking market analysis or advice from certified market strategists is important.
People betting on NVDA right now to become a $10tn market cap company are going to get taken to the cleaners when they realize the AI CAPEX cycle won’t grow to infinity and is instead cyclical.
how come there is no mention of valuations; how much of growth is already priced into nvidia?
I've got a few things to clear up and share some key details with you. My number breaks down like this:
First, plus 1215... Next, 403... And finally,
4030. I'll be waiting for your W.A message.
After so many years, ROCm is still shit compared to what Nvidia is offering. It is so annoying for consumers too.
AMD are a very well run company. However I think trying to catch NVD in AI is like trying to create a social media platform that competes with Facebook. NVD is taking the AI market and solidifying its position. If NVD have a mis step then AMD could catch.
Actual, thoughtful insight in TH-cam comments....
With a CEO that has the personality of a potato, you know this company's going into the dumpster.
We could be seeing that mistep soon.
Keep in mind back before the AI hype bubble Nvdia has had a long track record of lying cheating and scamming.
I don't think most people quite understand how deep SMCI and NVDAs ties go, not only do they have very close business ties but Huang and Liang are literal best friends. . I'm beginning to suspect why SMCI can't release numbers and their accountant just quit is because they have no real way of hiding the fact they were helping NVDA spoof their obviously fraudulent numbers. Lets not pretend NVDAs other customers aren't in on it as well.
Wouldn't be the first time NVDA has put out fraudulent numbers.
What kind of miss-step? What would that look like?
@MrChev-wq1jb
SMCI being outed for helping NVDA fake shipment numbers. I think that's what we are seeing with SMCI right now.
And these prices buying NVIDIA stock is the worst idea. Especially taking into consideration all the latest hints there is a problem with simple scaling. 🤯
I wanted to stop and thank-you for all your videos! You have helped me find some true gems. I have owned NVDA for 20 years (thus the intrust to start). But I never would have found PLTR in time for this years push. Again Thanks and glad to see you back up and running!
I'm so happy to hear that! Thanks for the kind words!
In conclusion, the Tickersymbolyou suggests holding AMD if already invested, but for investors with a strong focus on AI, Nvidia offers a more compelling growth story due to its market dominance and rapid innovation cycle.
Perhaps you should research the AMD Versal solution a bit more. It is an FPGA, ARM core & GPU merged into a single platform. FPGA's are superior in many ways compared to GPU's.
Low-Latency Customization: FPGAs, like those within the Versal, allow for custom data paths and parallelized logic that can be tailored precisely for specific low-latency tasks. Since the hardware can be configured to execute specific functions without needing to pass through layers of software, FPGAs typically provide lower latency than GPUs, which rely on general-purpose cores.
Deterministic Performance: The Versal platform, with its FPGA component, allows for highly predictable, deterministic performance, which is essential for latency-sensitive applications (e.g., real-time signal processing, high-frequency trading, or edge AI). Deterministic response times are often harder to achieve with GPUs, as their performance can vary based on workload and memory access patterns.
Real-Time Data Processing: For tasks requiring immediate, responsive processing, such as those in telecommunications or automotive applications, the Versal’s mix of FPGA, ARM cores, and AI engines can handle data streams and execute logic in real-time. In comparison, GPUs, despite their high throughput, typically introduce more latency due to their software stack and reliance on batching operations to maximize efficiency.
Direct Hardware Control: With the Versal, developers can set up data flows directly within the hardware, reducing the need for time-consuming data transfers or memory bottlenecks. This control over the hardware can be critical for latency-sensitive applications, allowing FPGAs to outperform GPUs in these scenarios.
Power Efficiency in Low-Latency Workloads: In latency-sensitive scenarios, the Versal solution can perform specific tasks more efficiently in terms of both power and speed, as it avoids the general-purpose, high-power approach of a GPU. Blackwell GPUs are optimized for sustained throughput over latency and often run best with a certain workload volume that compensates for their higher power requirements.
NVDA blackwell should be considered a general workhorse with massive parallelization abilities at the cost of latency and customization. AMD Versal is superior for latency sensitive applications, GPU's will never dominate this space.
AMD Versal will eat market share of Blackwell and for a lot of real time sensitive or edge products, will be the only option considered as Blackwell is weak in this area.
@@mcspeed07able seems like a solid justification , however at times stock movement is like fashion industry that follows trend (or) like movies with famous actors make box office than movies with good story content ... therefore from stock point of view if NVDIA falls rest will fall - if NVDIA moves up rest may follow ...
@@mavensept except time and time again we see the exact opposite, with Nvidia falling and AMD taking prominence. Yet the vice versa happens and we act like AMD is dead all of a sudden lol
@@mcspeed07able Doesn't matter what AMD has. The world has moved on. VHS sucked, but once it took hold of norms it didn't matter if Betamax was superior.
@@RajDeelish You are extremely wrong to use that example in this case. The Mega Cap tech companies will quickly move to whatever option is best and are actively working on custom chips with various partners. This industry can do a sharp turn at a moments notice so don't get complacent.
Having programable ASIC style chips from xilinx is a major game changer for AMD it was definitely a positive acquisition for them and put INTC on notice. Nvidia Mellanox acquisition was a huge game changer for them as well.
What matters is Tetaflops per watt, because it affects the price of cooling and electrucity per computation performance.
Please update about super micro computer. Are you still bullish or moved on?
On it
@@TickerSymbolYOU LMAO🤣🤣🤣
Absolutely on it, I don’t like money.
SMCI auditor resigned - risk is super high now - SELL
Delisted soon 😂
Do you really think META, GOOG, TSLA, MSFT, OpenAI, AMZN, APPL are going to let NVDA be worth more than all of them combined? I think over time these companies will a) Create their own chip b) help AMD (especially in software area to catch up to CUDA) compete with NVDA.
apple and intel are trying to make their own chips, and amd is trying to catch up to where nvidia is RIGHT NOW - while nvidia is already building the next big thing, nvidia gambled 10 years ago on AI by building up their software and the entire architecture around AI - not just the chip but all the parts the chip relies on, the b100/200 isn't just gpu's - it was rebuilt from the ground up & inside out to handle the better gpu's.
This is all true uintill the AI bubble holds. By now having a 150000 GPU cluster that powers an LLM that can barely give you some resemblance of thinking is pretty underwelming.
AMD also won’t tank from an AI bubble burst if one happens, especially compared to how much NVIDIA would due to it being the lions share of their company
AMD has created a niche for itself by offering "comparable" processing power at a price point below its largest competitors... AMD will be fine...
Datacenters are a sleeper market that people forget about. We forget that the internet lives on physical computers, thats what datacenters are. They are already huge and will be part of out infrastructure indefinitely. Targeting those in whatever way seems valuable will be a good investment imo
You channel is like comfort food.
I tend to get your recomended videos after I make a switch in stocks.
I got rid of shares of apple since I don't think they are innovating any more.
I found Nvidia, AMD, FLEX, and Tesla, at the start of the year being a good choice.
I already had google, amazon, microsoft, so these were a good swap for apple.
This year was great, but thanks making me keep a closer eye on AMD.
I've held both stocks for the past 5 years. No plans to sell either for at least another 5.
Patience makes millionaires. 💪
Hi Alex! Please make a video comparing Vistra, Constellation and Talen Energy companies. I'm sure your viewers would love that! Powering AI, and EV cars of the future. The stocks have already taken off. What's your take on all of this! Especially on Talen! Too late to get in?
Also I think for PE ratios it is 47 to 42 if I am not mistaken which means AMD stock is almost as highly valued as NVDA. I would be more into AMD stock if its PE was in the mid to low 30s especially when I couple that with all the other points you made on the video.
Look at forward P/E, not trailing P/E... AMD forward p/e 28.49 NVDA forward p/e 34.60... PEG ratio AMD (5yr .38) NVDA (5yr 1.05). We saw what happened to NVDA when they started realizing profits from data center sales... AMD is at the beginning of that same ramp now...
@@mr2d2sorry not the same. NVidia had no competition back then.
@@mr2d2@mr2d2 Good point, and honestly, I do not know as much about comparing forward P/E ratios but even with that said your point about the AMD ramp up is a good one to think about = POTENTIAL is significant for sure! I still believe NVDA has 4 quarters of uninterrupted ramp-up in all aspects of sales and the stocks, but I will keep some money in AMD stock also. I guess we will find out on November 20th if NVDA is still hitting it out of the park.
@@geolykos no idea what you’re trying to say… you think nvidia is overvalued now because they have competition from amd?
@@mr2d2 you said it’s the same as NVidia when they started realising profits from data center sales. it’s not the same. NVidia had no competition. AMD has to compete with a superior competitor for these sales.
Your SMCI stock pump did not age well.
How about every other company on the list in this video?
no one knew what SMCI was doing. He’s not pumping he’s simply give you info. If you buy or not is on you not him.
Hard to predict SMCI to cheat with their accounting manipulation.
Sounds like you lost a lot on SMCI stock. I did as well. About half of what I've invested into the company before I got out. Things were going great until we learned that SMCI cooked the books. No one knew what SMCI did until it was too late. I don't blame TS You. Motley Fool, in there Hidden Gems publications pushed five Chinese stocks. Three plummeted on the same day when it was learned that all five companies had cooked the books. Again, I lost a butt ton of money. I didn't Blame Motley Fool because they lost as much as I did. Even if I invest into Nvidia, the stock would have to more than double for me to reclaim my losses but that's exactly what I am going to do. Wish you well in your endeavor.
@@enigmaxd9853 says the person on YT mooching free videos and expecting 100% perfection.
They mentioned speed but they didn't mention energy requirements or TCO. Show me that clip then you'll have my attention.
Arm is not necessarily the architecture of tomorrow. They are alternatives that are still in development to x86, because no one could product x86 due to licensincing. X86 has really all the software support, which arm is catching up, tho its a long long eay to go. Performance of x 86 per mm and connectivity on epyc is still unmatched
🥳✌️🥳✌️Cool Bud! Glad you're safe.🙏
What about Vertiv, VRT? I've made a boatload with them in the last year but not sure if i should take my gains
We are definitely at the start of AI. Excited to see where these two companies will be in the next few years
@Ticker Symbol: YOU
why does your video feel so "lip-synched"? Do you extra voice over for your video? Or is it a slight delay between video and audio?
Feels so uncanny.
Hmmm I’m not noticing the delay but that’s probably what it is. I’ll watch out for that, thanks.
@@TickerSymbolYOU Now that I read this comment, I see it too, but I'm not sure if it's a psychological effect. If there is a delay it could be as little as a frame or two.
Nvidia has the software stack (Cuda) as the standard. Amd doesn't have access to that API (without a huge amount of effort (ROCM)), talent pool and technology. So their GPU is not competitive. Though their CPUs are very competitive in the server market.
I initially did the big ticket buy, 125k into SCHD, 75k TSLA, 25k VYM, 25K VUG. Now I'm dca buying roughly 2k every week of whatever is on sale, and looking to add more tech positions to my portfolio. I'm looking to hold long term 15 - 20 years, so hopefully my lump sum buy in doesn't bite me in the ass long term.
In the past month, my "unexciting" index funds provided me with over $6,000 in dividends, giving me the option to spend without selling shares. Currently, I've opted to reinvest the dividends to acquire additional index funds for future growth.
Nvdia is shifting to a strong software business...And no one is really focused on their future earnings on that front
I don't think you've ever worked with nvdia's "strong software"
@JG-nm9zk Just wait until the regular Wallstreet humans start talking about the software growth in revenue ... maybe you remember this ....I knew that GPU would be the brain of Ai computing 10 years before anyone was talking about it.....people were all saying it's a gaming business..lol
Why ARM? Risc 5 would seem to have more potential and is not encumbered.
You didnt mention Sound Wave, AMD's ARM-based APU set to release 2026. why did you not include the P/E ratio?
Thanks for all the excellent knowledge bro
Glad you’re safe. AMD has been counted out a lot in the past and Jensen once worked there. I think they’ll still be relevant.
It all comes down to cost. That's the bottom line for any business. How much does Bank of America want to spend on a chat-bot to answer customer questions? Less then the call center employees. Driver-less cars? They still can't get that to work. (Even on trains).
Would love to see a new video on SMCI since its current slide. Thinking this could possibly be a huge value opportunity for investors.
Keep an eye out for an update on that soon.
It’d be foolish to try to jump into SMCI right now. They’re on the road to Bankruptcy and delisting.
It's not about jumping into a stock -- it's about following up on stocks I've already covered
@@RedeemerBlood SMCI is toast. Its like Enron. Get out now before they go bankrupt and get de-listed.
Still don’t understand what Ai has done that justifies this oversaturated market. Id rather go with AMD since Nvidia can plateau.
You better compare first the gross margin of the two, then think about if you still sticking to AMD.
👍👍👍video makes lots of sense. Somewhat bullish on AMD but downgraded it? Thanks
Exactly right. Still bullish overall, but don't like how spread out they've become.
Thank you so much for the video, I like your quote " understand the science behind the stock".
The interconnect issue you mention is crucially important, it's not glamorous to talk about it, but it IS how we scale data centers.
Ethernet is simply superior in cost of implementation, universality and labor force specialization and support. AMD has actively collaborated with the group and is well positioned to leverage that scale and this leaves the entire industry wide open for their solutions. Nvidia preferred the thunderbolt approach. We all know how that ended up.
AMD and their approach means they WILL capture a larger share of the market. Its just simply going to happen.
Blackwell is based on H200 so it’s not wrong that she’s comparing MI325X to H200
For general computing, RISC V may be the winner in the future. ARM/Qualcomm IP issue got others looking another way. So will see how long ARM will run servers, Single Board Computers, Smartphones, etc. As for AI. NPU / AI ASICs on the same process node X nm will have better performance per watt compared to GPUs. So I could place a bet on Cerebras. I want to see more ASICs for general-purpose compression like ZSTD, Brotli, and others. Cast Inc. is a player in compression/decompression accelerators.
I hear what you’re saying. I would have stayed invested in AMD but I have so many more lucrative stocks to put the $ in…as well as I’m not convicted that Intel was a good decision. Thanks, Alex!
I’m glad they started adopting easy to remember names like MI325X. I mean in the old days you had Pentium and Pentium II.
There's no reason to make ARM chips in 2024. Instruction sets make up less than 1% of modern CPU's. All hardware designers are diversifying their compute units to offer more specialisation (i.e. larger instruction sets along RISC principles).
Energy efficiency of ARM chips can be fully explaioned through them usually being one processing node ahead and the focus on SoC integration, which is something AMD has a lot of experience with.
I really enjoyed it it's in plain English in a way that anybody can understand it and the difference that exists between AMD and Nvidia
Betting on the underdog is for horse racing, not stocks
I always keep hearing from Super Investors, it's always far better to be invested into the #1 clear leader. And not to chase after the #2 or #3. Who is the clear leader in the AI sector? We know.
Yep. I should take note to learn from the best more often. Lesson learned for 2025 and beyond!
Compare AMD market cap with Nvidia market cap. AMD has nothing to lose, while Nvidia has everything to lose. AMD is a steal at these prices imo
Yeah, and AMD will never be able to even come close to touching Intel, especially in the server room where the biggest profits are - that's what talking heads were saying just a few years ago. AMD has already announced (over a year ago) that they are developing ARM chips, and I would wager that their outside the box thinking will come up with market leading ARM chips in short order.
Dude we need this SMCI inside scoop!! You’re the only one I really trust to tell me what’s REALLY GOING ON!!
Releasing something as soon as I have something solid to say! I'm on it!
@ thanks man!
Good perspective on AI chip market. Sounds like typical nvda drop is in progress before next run up to nov. ER and blackwell release....
People forget desktop market is not a laptop market and it's not a corporate market.
If you want to invest you should care mostly about corporate market (AI, servers, networking, datacenters, FPGAs etc).
If you care about parts and their performance, look at hardware individually per platform and price range.
I wrote this because I am both a tech investor and I love new consumer tech too, just about tired of people fanboying over specific corporations and behaving like cavemen.
PSTG? Up 100% Jan - June, as of early Nov up 30% YTD. Will Pure Storage remain relevant as AI & super computing advance? More so?
and NTAP. These have both done well for me, but I don't really understand their products and how to think about them going forward.
Nvidia is working with Ionq on quantum and classical computing hybrids, not sure how AMD would overtake Nvidia if they’re not in this space at all.
Good presentation. However I think AMD AI revenue will grow as a percentage of sales going forward which will increase the overall growth rate of the company. Beside that the stock is selling as approximately 23 times non gap eps of 6 for 2025. Furthermore their gross margins will trend upward too. The issue with AMD is that people keeps comparing it with Nvidia the undisputed king of AI. It is a bit unfair. AMD was able to beat Intel with much less financial resources. It will be a different ball game with Nvida which as a very different and aggressive company. I think AMD could aim at 10-15% of the AI accelerator market within 2 years. Finally we can see that the accelerate product release of Nvidia has created some technical over heating issues that could help AMD. TIME WILL TELL. thks for the video again. Good work😊
Thank you so much for bring us such a high quality information. I have a question...are these superclusters still needed once the model has been trained? what happens to the hardware once training has been completed?
Training is never really "completed" -- models go through fine tuning and re-training as new data comes in while new models are constantly being built (either better/bigger general models or more application-specific ones)
@@TickerSymbolYOU Thank you, appreciate all the knowledge you share.
IT speacialists are recommending AMD sell their old cards as cheap as posible and prepare lunch RX 8800 series as low as $400 if AMD want to keep at least 25% of market share from nVidia-nPolosi.
How do you feel about Taiwan Semiconductor? It seems more poised to benefit from AI spending than both Nvidia and AMD long term because companies like Google, Apple, and Meta can and likely will design their own chips
I'm keen on AMD's upcoming APUs it'll be so cool...once my current desktop dies I'll be moving onto a MINI PC setup. building desktop rigs is becoming expensive each time
Compared to Nvidia, AMD have alot of room to grow and take market shares in the future.
That was my initial thinking too
I hold AMD for 2 years and only make $34k on compare NVDA also holding 2 years and making $397k
Years ago when AMD was on the edge of bankruptcy, I bought about 25,000 shares at $1.50, but I was chickened out and sold after it dip to bottom and went back up. However, I bought NVDA after that. In 2022 I bought some AMD at around $65. In late 2023 when Lisa gave guidance 2B when Nvidia could not made enough chips to sell, I said to myself this is not good for AMD ecosystem for the years to come, so I sold all AMD (sold TSLA, too bc Elon promised and not delivered) and added more NVDA shares. That amount of money has now been 3x while TSLA and AMD are going nowhere.
Nvidia is buying back shares. AMD is diluting shareholders with stock based compensation. Nvidia has stock based compensation as well but they are buying back more.
@@Sunil-he8xm i only own Nvidia so far, but i see the potentional of AMD in the upcoming years, but yes Nvidia buying back shares is really great.
With the free cash flow and ecosystem gap growing bigger, it's harder to catch up.
Why do you assume they are wanting to catch-up - horses for courses. You don´t have to be the biggest - to be the best. This is a US mindset - but the rest of the World we go for quality rather than pack em cheap - sell them high mentality.
Ok. How about "harder to compete?" When it comes to accelerated computing and AI, harder to compete also means demise because AI progresses a hundred times if not thousands time faster than technology in the past. It's not like personal computing and games where you just bring some little chips to the table. It needs the whole stacks, platform, ecosystem, and partnerships. I have not even gone into arguing about when physical AI comes, AMD will have nothing to compete with Nvidia's Omniverse. I have subscribed to 2-minute-papers TH-cam channel for almost 10 years. Never heard of any AI paper from AMD while papers from Nvidia presented almost every week. All the billionaires and giant companies that go after Nvidia's chips and technology send me a message. Don't underestimate billionaires bc they are billionaires for a reason.
@@notbaltic280 The so called investors in the US only look at quarter to quarter. Most are just traders not real investors. We have been conditioned by the talking heads and the so call analysts that are by and large often wrong. It seems buy and hold is dead, which really hurts any real investor at least short term. I talked to a investor relation person one time and he told me they prefer institutional investors over retail. Retail is always the most irrational, I assume because institutional is not using their own money.
I'm concerned for the long-term future of x86.
Apple's custom ARM chips are very good and it's only a matter of time until there's a huge disruption to the Windows PC market.
The Mediatek+Nvidia partnership is interesting but they may not be working on high-end PC SOC's, just mobile/laptop SOC's. It'll likely be better than Qualcomm's recent attempt at ARM on Windows.
Completely agree. I think the future of processing isn't necessarily on x86
Is couldn't believe it. Went ahead and bought a ton load of INTC.
If MI325X is "good enough" it can potentially not matter if Rubin is significantly better and AMD could find a healthy market with MI325X.
NVIDiA covers all aspects of the AI revolution. You can’t become involved in AI as a company and not include NVIDiA in the process! They are the dominant company in leading the AI revolution! It is a big part of my portfolio and will be the reason why I retire early! Good luck to everyone invested in NVIDiA!
Without CUDA alternative and stronger drivers development AMD has little chance against Nvidia in GPU and AI market. And this is not a problem since companies are closely connected.
It would’ve been also helpful if you included your thoughts on AMD’s acquisition of ZT Systems. Personally, this is a great deal that helps AMD gain more partnerships with hyperscalers, and AI infrastructure. Any thoughts?
Nobody talks about NPUs, neuromorphe chips. Why?
I really enjoy your content and look forward to future videos.
Is AMD supply-chained to TSMC, feeding off whatever production capacity Nvidia has not consumed? Perhaps AMD should strengthen that strategic partnership with Intel to include foundry contracts.
$3.5 T vs $240B, not in the same league.
Think you overlook manufacturing
Welcome back bud. Glad you're safe
I would suggest that AMD never had a headwind in supply, since both NVIDIA and AMD are reliant on TSMC, and to smaller suppliers; but there's 2 very important points that you're missing in your analysis. BOTH AMD and NVIDIA will be moving to a yearly release cadence. Also, because no one but AMD and Intel have X86 licenses (I believe VIA is the 3rd), they have to release ARM. But, it would be a mistake to think that neither company could release ARM processors themselves. They in fact, do; just not bread and butter chips
Most of my portfolio is in both of these and I'm winning huge!!!! I mean HUGE! Average yearly increase is about 120% yearly over 10 yrs-that's as far as I can get an average.
I have been investing for years. Unfortunately I missed the opportunity to jump off at $200, I should have used the AI hype.
They simply cannot deliver like NV. That has become clear over the years. No question, AMD is a good company, they deliver well. Intel has done well, but unfortunately the profit is only made in AI and not PC or data center CPUs.
The lag in software quality and support does the rest. Everything is hopeful, but it takes years to prevail. NV has simply beaten them so clearly, year after year.
Don't buy a single stock, just buy the entire sector. Because you cannot verifiably predict which stock will win the AI race, but for sure of of them will.
Even people working directly in this industry cannot predict what will happen! When I worked with Nvidia hardware for AI research, I was sure it would never make it to the top of the AI hardware game, just because it's so full of glitches and errors. Yet, look where we are.
What happens to NVIDIA stock if SMCI gets delisted?
Probably not much. I imagine Nvidia has partners/installers willing to pick up the slack and then some. I'd imagine a minor slowdown in short-term sales and that's it.
One day those billions of investment in AI has to start making a "return of investment". I don't see where are they going to get that return from. Unless this is more of a "infastructure investment" in a conventional sense.
"Skynet" is just over the horizon and it might be here sooner than we thought.
Are these chips using photonic back plates?
Alex what to you think the AI PC market will be and is this a big market opportunity for AMD? Not so much for consumers but enterprise.
I think all markets crash except maybe commodities. Be careful folks. ❤