Everything is sold out.

H100s… gone.

H200s… gone.

Blackwell… they can’t make enough.

Big cloud players are literally renting capacity from each other just to keep up.

That’s the reality of Nvidia’s business right now.

Yet some investors still call it a bubble.

They said it last quarter. The quarter before that. And the quarter before that.

They’ve been wrong every time.

Because Nvidia isn’t about what happens in a single quarter.

It’s about the scale of what comes next.

Here are my three biggest takeaways from Nvidia’s earnings.

Takeaway #1 – NVDA Just Got Cheaper

When people look at Nvidia’s stock price, they see a big number.

$180 per share. Up 12x from when ChatGPT was released.

But what matters isn’t the sticker price.

It’s how the valuation stacks up against earnings.

And here’s the thing: earnings keep climbing.

Since the launch of ChatGPT in late 2022, earnings are up 17x!

And they’re only expected to grow higher.

Over the last 90 days, expectations for this year’s EPS jumped from $4.27 to $4.45.

Next year’s numbers rose even more – from $5.73 to $6.18.

This kind of steady upward revision is one of my favorite setups in the market.

When investors see “expensive,” but the fundamentals are improving so quickly that the stock is secretly getting cheaper.

Nvidia fits this pattern perfectly.

Takeaway #2 – Growth, Growth, Growth

This quarter, Nvidia CEO Jensen Huang dialed up expectations to another level.

Previously, he saw $1 trillion in AI infrastructure spend through the end of the decade.

Now he’s calling for $3 trillion to $4 trillion.

That’s not just growth. That’s exponential growth.

We’re still in the early innings.

Right now is the training phase of building these massive AI models.

Next comes inference, or reasoning, where the models actually get deployed at scale.

And after that comes physical AI: robotics and autonomous systems that require vastly more compute power than anything we’ve seen.

As Jensen explained on Wednesday:

“And the amount of computation necessary for [reasoning] models could be 100x, 1,000x and potentially even more.”

And…

“Robotic applications require exponentially more compute on the device and in infrastructure, representing a significant long-term demand driver for our data center platform.”

No wonder Wedbush just estimated that demand for Nvidia’s newest chips stand at 10:1.

Nvidia is experiencing demand at a scale that we’ve never seen before.

And it isn’t slowing down.

Takeaway #3 – China Would Be the Cherry on Top

If there’s one wildcard that could make Nvidia’s growth story even more explosive – it’s China.

Jensen estimated the China market represents a $50 billion opportunity this year alone if Nvidia could fully address it with competitive products.

And he believes it could grow 50% annually from there.

Regardless of what you may hear, here’s the reality: Chinese AI developers both want and need Nvidia’s GPUs – regardless of what their government says.

Nvidia is best in class and if China wants to build the best AI models in the world, they have to use the best hardware in the world.

The country is also home to about half of the world’s AI researchers.

Many of the leading open-source models are being developed there.

If Nvidia can sell into China again, even partially, it could add rocket fuel to a story that it's already dominating.

Here’s the Bottom Line

Nvidia’s earnings once again proved the skeptics wrong.

Yes, the stock traded flat after the announcement.

But underneath the surface, the setup keeps getting better.

Earnings estimates are rising.

Growth prospects are accelerating.

And a potential China reentry looms as a massive bonus.

This is not a bubble stock waiting to crash.

It’s the backbone of a once-in-a-generation technology shift.

Own it.