Your entire retirement savings is probably already pretty much only in NVDA if you are in a 401k…
Please. I live is SOCIALIST Canada. I have an RRSP like a civilized person.
2 Likes
My main 401k fund has NVDA at 2.6% of the funds holdings. Apple, Microsoft and Amazon all ahead of that.
sounds like SPY or some other S&P etf
♪ I got my first real six-string ♫
♪ Bought it at the Five and Dime ♫
♪ Got in TSLA at Three-Hundred-and-Eight♫
♪ Now it’s worth One-Hundred-Fifty-Nine ♫
6 Likes
I kinda hate these things because it’s yeah what the fuck but the reason these companies get these breaks is because they offset there taxes against research and development. You can almost certainly argue all extremes from that’s worth shit to that’s worth more than the taxes they would have paid. But it certainly isn’t worth shit which is what these click bait headlines imply.
It annoys me almost as much when maga fuckwits really believe we are giving (insert any shitty country) aid out the kindness of our hearts.
1 Like
“best chip maker for AI” is kinda meaningless. we are so much in the beginning of almost all of GPU and AI research that’s going to come, it’s impossible to project that superiority forward for 10 or even 5 years.
like Intel was the best chip maker for desktops for a solid decade and a half, now basically any phone maker and any computing provider can outsource the design and manufacturing of CPUs specifically for what they are doing with a few billion dollar investment. the investment itself isn’t much of a barrier anymore. the barrier is figuring out how to sell the product of that investment for a few billion profit.
S&P isn’t even the Mag7 anymore, it’s just semis
FSKAX, Fidelity Dow Jones clone basically. By far the lowest expense ratio offered in our plan
1 Like
Right, and that is where the volatility is going to come from, Nvidia has been best at it for long enough for them to grow over 200% since last year. Everyone who is buying chips from them wants cheaper chips but there are only a limited number of players who can even start to compete with them.
It’s not only money, it takes time to be able to build chips that are going to compete so if you’re an Intel or AMD could do it but it will take time, otherwise it will take a lot of time…
I put all my 2024 contributions into international (SWISX), VDE and VFH. Felt like I already had enough exposure to the four companies you mentioned.
how much longer do you think gpu revenues going to 3x next year? how much longer do you think total GPU deployments are going to double YoY? how expensive do you think the GPU equivalent of a mega-FLOP going to be YoY? how many of those FLOPS are going to be from a competitor?
i think if you put a probability distribution on those questions, NVIDIA is likely going to miss compared to the year they just had, still going to grow market cap obviously, but i would bet the field here, rather than NVIDIA. they are not exactly swimming in cash, and it’s not a given they will fulfill all of current demand.
I doubt NIVDIA can keep growing at this pace, sounds like they are already selling everything they can make. I don’t see GPU sales slowing down for a while, there are going to be too many start-ups that need servers for that to happen. No idea on FLOPS but AMD and NVIDIA are on yearly release cycles so the chips are going to keep getting faster and with current demand there isn’t much reason to make them cheaper. NVIDIA has the advantage of their CUDA software which will take more time for a competitor to match, if any are even trying.
I doubt NVIDIA has the capability of being the Intel of AI chips, no one wants another Intel and there is enough money in the AI market to bring in competition. But I think there is still room for growth in their stock and I’m going to hold to my 1 share for a while…
1 Like
those arguments are fine and logical, and the counterpoints are also relatively straightforward. for one: i think a lot of AI workloads on GPUs are going to be turned off and replaced by more sensible and more efficient workloads. so more websites and applications with weather and climate models, and fewer websites where you can generate several nude versions of a scene from oppenheimer.
another surety: the expensive thing in computing power is scale. it’s already cheaper to curate the input down and make the AI models more specialized and more precise. and it is already more than 2x more expensive to feed models 2x data in pursuit of generality. because of “double every 18 months” philosophy, everyone is eventually going to run into physical and financial limits, just a matter of has it already happened last year, or will it happen this year or the next.
the other trend we are going to see is likely increased competition from nonGPU makers. every country wants to get ahead of next chip and semiconductor shortage, they’ll be subsidizing mini taiwans domestically. and GPUs play into that, because the process to make GPUs consumes 100x+ of the silicon wafer than a nonGPU chip. it won’t kill off GPU business of course, but i see future growth a lot closer to 30% than to 200%.
just my hottake.
1 Like
Can someone just tell me when the fuck to sell my NVDA? I bought it like a donk with no specific plan, so I sure don’t have an exit point in mind.
The traditional answer is that if you wouldn’t buy it now, don’t hold it now. You probably owe lots of taxes on it, but that’s a pretty good problem to have.
1 Like
The taxes are what has kept me from selling it. I’m sure this will also be my downfall.
I can’t TLH enough to offset a significant amount.
I know it will be more painful than this sounds, but if you’re already in long-term gains territory just take the hit. (I assume it’s 15% federal for you, plus potentially state and net investment tax.)
I’ve held for over a decade. It’s gonna be brutal. I know I shouldn’t complain.