For what it's worth, I do not believe nVidia is the future. I'm not predicting they will go out of business tomorrow. I just think they've lost a whole lot of mind share in GPU and their dominance in AI is not as strong as it may seem. They are limited by TSMC capacity, AMD is ramping quickly, and Intel is just dumb enough to stop working on everything that matters and bet their future on AI. Everyone with an NPU will be Hell bent on global domination. nVidia used to dominate but I believe those days are over, based on nVidia internal corporate culture. Again, I'm not saying Intel will go all in on AI, just that I wouldn't put it past them.
Bold words indeed. I will give you that it was inevitable that competition would begin to really rally around taking down the king, but I am not convinced it is that case - at least not yet. I am seeing reports about 1-bit AI models on CPUs and using lesser hardware on DeepSeek-esque models. The moat is not as deep and wide as it once was, and I think the Rubin rollout and orders will be a big tell. I also think you are very right about execution and corporate culture, at least as it is right now. For all we know, Jensen might be raging and on a redemption rampage due to Blackwell and its woes. He's that kind of CEO. Again, Rubin is the key for me. As for my investing strategy, I will compare NVDA to VUG. As long as it outperforms VUG in one year increments, I will keep fully invested. If another company emerges as the next Nvidia, I'll jump on that, but until then, all it has to do is outperform VUG. Can't ask for more than that.
Quite a bit of Blackwell product (RTX 50X0 series) volume is well up in the channel. It just started this week. I wonder if this means AI demand has been satisfied.
It sounds like xAI's Colossus GPUs have been fulfilled. That's probably why there is more nVidia product in the channel.
Word from the channel is that nVidia midrange GPUs are not selling out at MSRP. Wild, if true. AMD GPUs are selling out above MSRP.
Too little VRAM, too much FU to gamers. Nvidia deserves the bad PR this generation. Blackwell has not been a good launch in either consumer or datacenter realms.
Wow, Nvidia playing dirty now: Aww, I'm just kidding, they always play dirty. That's why they are #1. But seriously though, Jensen is a big douche. Just look at the jackets!!! Still on the Mount Rushmore of CEOs, but a huge douche. Like Steve Jobs. Now go make me more money to be able to afford your GPUs, you big douche!
nVidia's bad will has definitely caused nVidia to lose massive mind share to AMD. AMD hasn't tried to compete on the high end of gaming GPUs, choosing to just leave that market for nVidia. I believe that has been a reasonably smart move on AMD's part. The new politics have increased the value to AMD to take a shot at the high end market. They are working on it. nVidia dominates the performance market with a significant performance lead but their silicone has little to no lead. To achieve what they have, nVidia silicone requires massive power. AMD doesn't have any sort of lead over nVidia, either. The fastest AMD GPU consume close to 500 Watts at their limit. To compete, they will either need to use multiple dies which are slightly tuned for efficiency. It won't take much of a performance hit knock their GPU back to 300 Watt peak, and that will open the door for a second die. Or, they could make a single, larger, die and see how fast they can go at 600W TDP. The biggest expense will be the new cooling solution for the 600 Watt part. Fortunately, nVidia has created an ecosystem of power supplies and mindshare for GPU power consumption at this level. I'm 75% confident AMD will attempt a 5090 class competitor. They are talking about it. The question is, how much damage has nVidia done to their franchise value in gaming? It may not be as bad as the YouTube crowd think. Intel has done heinous things to their customer base over the years and didn't lose mind share until only recently. I suspect nVidia can regain significant market share with Blackwell Ultra, later this year. They might regain nearly all of it. The issue I see is that nVidia has accomplished 4K, high refresh, gaming. While imperfect, the 5090 is close. They need Blackwell Ultra to perform at a similar level or slightly above, while consuming considerably less power. If they could achieve a 500 Watt, 4K, GPU with significant driver improvements, they could make it extremely difficult for AMD to gain further market share.
Look out Intel and AMD, Nvidia is looking to take it all! NVIDIA's Arm-Based Gaming SoC to Debut in Alienware Laptops https://www.techpowerup.com/337574/nvidias-arm-based-gaming-soc-to-debut-in-alienware-laptops As an aside, x86 Windows days are numbered. Good riddance. Microsoft has turned Windows into an absolute mess. Linux is the future. Microsoft needs to own that fact.
The N1X is scheduled to be released toward the end of the year. I don't expect a lot from this project. As best I can tell, the work environment at nVidia is toxic. It was OK while money was flowing like water but it was never ideal. Toxicity is an impediment to the creation of bug free software. I hope the drivers, HAL, and microcode is relatively bug free but would not bet on it. If they can get it sorted, and I hope they can, the N1X seems to be targeted as an AI front end. Perhaps AI portable terminals and devices? The N1X is spec'ed to have 8 ARM cores and a whole bunch of GPU cores. It won't be a threat to AMD mainline CPUs.
Nobody sees it yet but nVidia is in one hell of a jam. This ship will sink more quickly than Intel but they can pretend to be on top of the world for another year or two. I watched an interview with an ex-nVidia employee and he said the environment is crazy toxic. Nobody wants to be there. For years, that wasn't an issue because they were all becoming multi-millionaires from the employee stock programs. So, they worked their butts off. Now, the money is starting to dry up and the number of people who will work like animals along with it. He is just one guy but there isn't a lot of forward looking good news coming out of nVidia. At the same time, Jensen Huang now sounds like more of a cult leader than a CEO. He says some far out stuff. It sounds like he lives in a thick bubble.
I've heard that the issue is that people are so rich, that they've lost the incentive to work. IIRC, the policy at NVDA is a 5-year look back on ESPP, which means the moment you join the company you are a millionaire. If people really thing NVDA is done, then go and short the stock. You'll become very rich!
Not people. Me. I don't short stock and don't plan to start. If you look back to February of last year in this thread, you will se me predicting AMD will come close to catching nVidia. That was based on AMD executing well while nVidia's work force being fat and unmotivated. Look at the push back I got. It wasn't just here. It was everywhere. Everyone knew nVidia was going to pull out an even wider lead. Now here we are. AMD is in the ball park. The tortoise has nearly caught up with the hare. Even in AI, AMD has a plausible chance to catch and pass. If AMD eats nVidia's AI lunch, nVidia is dead.
I suspect your contradictory nature is based on a distaste for what I am writing. The thrust of my point is, the speculative hoards are idiots. Jensen Huang seems to have partially lost his grip on reality. He may still be a good CEO but I have some concerns. The workforce is demotivated. Blackwell was a very weak refresh. Blackwell is more powerful than Hopper for AI scaling but not more efficient. Hardware gains were less than clock increases which came from process node stepping. The primary gains for the Blackwell platform came from the software side. Meanwhile, gamers have left nVidia in droves to run AMD cards. Now that nVidia has a bit more product in the channel, we can see some models are not selling out. This, even at relatively low product availability levels. AMD has gone from 8% to nearly 50% in the last couple of months. That's sales, not the installed base. Speculators who know nothing about the product or the company think nVidia is great. I'm trying to dive deeper than simply cheering for a winner. Varying opinions help us all. Why don't you set aside your contradictions and share your view on nVidia, where it is, and where it's going? It might be helpful to the rest of us.
No? I'm just pointing out a contradiction in your narrative. Now, instead of clarifying, you're attacking my motives. I find that generally, markets get things right. You've offered some reasons as to why the market might be overvaluing NVDA nad undervaluing AMD, but at the same time, as I pointed out, you made a bad call in February 2024 (and somehow warped that into being a good call!) I don't see any evidence of this besides an article on sales from a certain part of Germany. Seems like Nvidia still has 80-95% of the market. My predictions on chip stocks haven't been that great (although I'd argue that for AMD, ultimately it hasn't been that big of a deal), so I'm inclined to lean into the market on this one.
I called it exactly correct. There is no honest way to throw shade on those comments. Look at the post/edit date. They haven't been touched. I'm the first one on this web site to point out where I've been wrong, however, I simply wasn't wrong here. You clearly will contradict anything which does not support your point of view. nVidia currently has a lead. It's extremely minor in GPUs. It's major in AI. AMD is showing big strides in AI. AMD efficiency appears to be higher than nVidia but their platform doesn't scale as well, even with the huge advances they've shown in the last 60 days. nVidia exploded, rightly, on news of 0.5T federal investment in AI. That's today. Tomorrow is another story and what I've been talking about, all along. Time will determine the winner, as always. I expect that 18 months from now, if my hunch of AMD gaining parity with nVidia (or a slight edge) comes to pass, you will again pretend I was entirely wrong in this post. It's all here for people to judge as they see fit.
