Great write-up on Nvidia Alex. Being an Nvidia shareholder since 2018, I appreciate the fresh take on the company. I agree that the stock has been priced for perfection for years and that a slowdown would mean a massive drawback in the stock price. It isn't an if, but a when. Like yourself, I'm betting that AMD takes more market share over the mid to long-term; I am currently in the middle of reviewing my AMD deep dive as I type this response. Their chiplets and inference narratives are compelling and I have no doubt Lisa Su, who has seen a 100X+ of the stock price in her tenure, is the one to do it.
Great read! Very reflected view on the economic moat of the business. Nvidia is forced to keep pace with the technological development and to stay ahead of competitors. Sounds normal for all businesses in the market but if your "brand" does not manage to bind customers long-term, they have a difficult race to win and so far the market prices Nvidia as a winner long-term.
Welcome back! Really enjoyed this piece – the "great business vs. great investment" framework is exactly the right lens to use here. And I appreciate the intellectual honesty in the "I Could Be Wrong" section.
That said, I'm still holding my NVIDIA position (via eToro), and I want to respectfully push back on the expectations framework, which I think is your strongest argument.
On the Expectations Game:
You're right that the bar is incredibly high – anything short of perfection disappoints. But here's my counter: hasn't this been true for the last 2+ years?
In early 2024, everyone said "the easy money is made" and "expectations are too high." Yet NVIDIA kept delivering the "impossible" quarter after quarter. The stock tripled from there. The market kept saying "this can't continue" and NVIDIA kept proving it could.
So the real question isn't whether expectations are high – it's whether the underlying TAM and execution capability can continue to surprise to the upside. And I'm not sure we're giving enough credit to how early we might be in the AI infrastructure build-out.
The Competition Argument:
AMD's MI400 pricing at 50% discount sounds compelling on paper, but we've seen this movie before. AMD has had "NVIDIA killers" in the past, yet the CUDA moat and software ecosystem keep customers locked in despite higher prices. It's not just about compute performance – it's about the entire developer stack, optimized libraries, and tooling that took 15+ years to build.
Yes, hyperscalers are building in-house chips (Trainium, TPUs, MTIA), but could these be complementary rather than replacement? For the most demanding frontier model training and inference at scale, NVIDIA might remain the gold standard while in-house chips handle more commoditized workloads.
The Positioning Thesis:
This one intrigues me most. "Everyone owns it" feels intuitive, but is retail/institutional positioning actually that saturated compared to previous mega-cap winners at similar stages? And what about the international flows just beginning (sovereign AI initiatives in Middle East, Asia, Europe)?
Also – and this is the growth investor in me talking – if NVIDIA continues to execute and the AI revolution expands into enterprise, edge computing, and autonomous systems, couldn't new capital keep flowing in even with current positioning?
My Bottom Line:
I totally get the risk/reward calculus at $192, and you may very well be right to de-risk here. But for me, the asymmetry works the other way: the TAM expansion potential and NVIDIA's execution track record make me willing to ride through volatility.
Curious: what would make you reconsider? What would NVIDIA need to demonstrate (new product cycles, new verticals, sustained margins despite competition) for you to get back in?
Looking forward to the discussion – and thanks for the thought-provoking analysis!
The circular financing in this space is also very 'interesting'. Nvidia just recently pulled another one.
When one of them gets into bigger trouble, it could also have big impact on others. I'm watching with the interest...
That's right! Will be interesting
Great write-up on Nvidia Alex. Being an Nvidia shareholder since 2018, I appreciate the fresh take on the company. I agree that the stock has been priced for perfection for years and that a slowdown would mean a massive drawback in the stock price. It isn't an if, but a when. Like yourself, I'm betting that AMD takes more market share over the mid to long-term; I am currently in the middle of reviewing my AMD deep dive as I type this response. Their chiplets and inference narratives are compelling and I have no doubt Lisa Su, who has seen a 100X+ of the stock price in her tenure, is the one to do it.
Totally agree Gannon!
Great read! Very reflected view on the economic moat of the business. Nvidia is forced to keep pace with the technological development and to stay ahead of competitors. Sounds normal for all businesses in the market but if your "brand" does not manage to bind customers long-term, they have a difficult race to win and so far the market prices Nvidia as a winner long-term.
Welcome back! Really enjoyed this piece – the "great business vs. great investment" framework is exactly the right lens to use here. And I appreciate the intellectual honesty in the "I Could Be Wrong" section.
That said, I'm still holding my NVIDIA position (via eToro), and I want to respectfully push back on the expectations framework, which I think is your strongest argument.
On the Expectations Game:
You're right that the bar is incredibly high – anything short of perfection disappoints. But here's my counter: hasn't this been true for the last 2+ years?
In early 2024, everyone said "the easy money is made" and "expectations are too high." Yet NVIDIA kept delivering the "impossible" quarter after quarter. The stock tripled from there. The market kept saying "this can't continue" and NVIDIA kept proving it could.
So the real question isn't whether expectations are high – it's whether the underlying TAM and execution capability can continue to surprise to the upside. And I'm not sure we're giving enough credit to how early we might be in the AI infrastructure build-out.
The Competition Argument:
AMD's MI400 pricing at 50% discount sounds compelling on paper, but we've seen this movie before. AMD has had "NVIDIA killers" in the past, yet the CUDA moat and software ecosystem keep customers locked in despite higher prices. It's not just about compute performance – it's about the entire developer stack, optimized libraries, and tooling that took 15+ years to build.
Yes, hyperscalers are building in-house chips (Trainium, TPUs, MTIA), but could these be complementary rather than replacement? For the most demanding frontier model training and inference at scale, NVIDIA might remain the gold standard while in-house chips handle more commoditized workloads.
The Positioning Thesis:
This one intrigues me most. "Everyone owns it" feels intuitive, but is retail/institutional positioning actually that saturated compared to previous mega-cap winners at similar stages? And what about the international flows just beginning (sovereign AI initiatives in Middle East, Asia, Europe)?
Also – and this is the growth investor in me talking – if NVIDIA continues to execute and the AI revolution expands into enterprise, edge computing, and autonomous systems, couldn't new capital keep flowing in even with current positioning?
My Bottom Line:
I totally get the risk/reward calculus at $192, and you may very well be right to de-risk here. But for me, the asymmetry works the other way: the TAM expansion potential and NVIDIA's execution track record make me willing to ride through volatility.
Curious: what would make you reconsider? What would NVIDIA need to demonstrate (new product cycles, new verticals, sustained margins despite competition) for you to get back in?
Looking forward to the discussion – and thanks for the thought-provoking analysis!