Nvidia fundamentals 'off the charts' but full positioning weighs on stock (with Jonathan Ferro, Lisa Abramowicz)

3/19/2026 6:22:02 PM	short term cautious down	5 days later	+1.11%	-0.55%
10/28/2025 3:42:05 PM	long term up	61 days later	-2.28%	-2.28%
9/3/2025 5:31:38 PM	short term up	5 days later	+0.88%	+0.88%

When you say 'noise' around Nvidia, what are you talking about?

Jonathan Ferro

Vivek Arya

There is fundamental and then there is positioning. Fundamentals are off the charts: sales growing over 80%, third straight quarter of acceleration, gross margins holding at 75%. But positioning is full because 8 out of 10 investors own it.

They raised their dividend by a factor of 25, so the dividend yield is higher than Apple's. The stock has outperformed the S&P 500 over any 6 or 12 month period.

What is the biggest counterpoint to the thesis that hyperscaler custom chips will displace Nvidia?

Jonathan Ferro

Vivek Arya

Hyperscalers use custom chips for internal workloads, but for public cloud workloads requiring flexibility, they use Nvidia GPUs because it's a plug-and-play, exceptionally flexible architecture. Nvidia is keeping 70-80% of the value of this industry.

Even if the industry grows to $1 trillion in sales, 30% of that is still a lot of money. The public cloud, neo-cloud, and sovereign on-premise enterprise will still prefer Nvidia.

Will Nvidia be pressured to lower costs to compete with alternatives?

Lisa Abramowicz

Vivek Arya

There is always pressure, but customers care about price, performance, and reliability of supply. Nvidia's full-stack optimization (silicon, hardware, software) and ease of deployment are big advantages. On a cost-per-token basis, we continue to see deflation.

Every new generation of Nvidia products sees content go up by 50-60% and performance go up by a factor of 3-10 times. That's what customers care for.

What should we look for in terms of ongoing adoption with the potential for $4-5 trillion in spend every year?

Lisa Abramowicz

Vivek Arya

There is a race between public hyperscalers and private companies like OpenAI and Anthropic. They are pushing the ecosystem to build out this intelligence layer as quickly as possible in a supply-constrained industry. This is a global race that can continue until the end of the decade.

The industry is constrained in substrates, memory, optics, and CPUs. The build out of this global intelligence layer is not going to stop anytime soon.

SignalTube