Yesterday, we talked about Nvidia’s $68 billion quarter.  The chip supply wall has finally broken over the past two quarters.  As such, data center revenue growth is back.  

But if Nvidia is to fulfill on the demand Jensen has said is on the books for the next four quarters, global manufacturing capacity will have to have another step-change higher. 

If that happens, Nvidia would likely have around $130 billion revenue quarter by the end of this calendar year.  It would be growing at a triple-digit rate again (year-over-year).  Apply the current 60%+ net income margin to that size of revenue, and the stock (at a $4.5 trillion market cap) would look very cheap (a forward P/E in the low-to-mid teens)!

If the chip supply bottleneck has truly been broken, then Jensen’s formula of “compute equals revenue” (which fuels more compute which fuels more revenue) become constrained only by the physical stuff — scarce inputs like electricity, uranium, copper, fiber optic glass, land.

With this in mind, let’s talk about a few other comments of note, from Jensen’s earnings call yesterday.

First, he said “every data center is power constrained.”  And with that, he says data centers have to carefully choose the architecture (the AI systems) that maximize “performance per watt.”  He says Nvidia wins on that front, “unquestionably.”      

So, access to power determines whether Jensen’s AI formula (compute equals revenue) continues to work.

And that’s why Elon has said in the age of abundance, power/energy will be the true currency.  

The more abundant intelligence becomes, the more valuable the scarce inputs become. 

Let’s talk about Nvidia earnings.

In my note yesterday, we talked about the 2026 revenue pipeline — half-a-trillion-dollars worth (that’s 2026, calendar year).

Jensen revealed last October that the demand was on the books already.  But fulfilling it was/is a different story.  It would take significant expansion of global manufacturing capacity.

With that, we had evidence in Q3 that Nvidia’s growth was back (after seven quarters of stagnant growth) — that supply was opening up.

And we got more confirmation of that today, from Q4.

Here’s what the quarterly revenue change looks like now.

That said, to fulfill on Jensen’s half-a-trillion-dollar (datacenter revenue) target by the end of the year, manufacturing capacity will have to have another step-change — and soon.

But, what’s rapidly carrying more of the growth load for Nvidia’s data center revenue is networking.

Compute (Nvidia’s GPUs) gets all of the attention.  But networking growth is exploding.  It did $11 billion last quarter — up 34% quarter-over-quarter, and 3.5x year-over-year.  Nvidia is now (maybe) the largest ethernet networking company in the world.

Keep in mind, this is all Q4, prior to the game changer on February 5th.

What happened on Feb. 5th?

On that single day, two of the most important AI companies in the world released models that represented a big leap in model capabilities.

Anthropic released Claude Opus 4.6, with autonomous agent teams that can manage entire workflows in parallel. Not answering questions. Doing the work.

Twenty minutes later, OpenAI fired back with GPT-5.3-Codex — a model that helped build itself.

And then, almost overnight, an open-source personal AI agent called OpenClaw went viral. People began running autonomous AI assistants on their laptops, around the clock, with access to their files, their browsers, their messaging platforms.

This is the moment AI crossed from “AI can think” to “AI can act.”

On that note, on the Nvidia call, Jensen proclaimed that the ChatGPT moment of Agentic AI has arrived.

What else did he say?

He said this line over and over:  “compute equals revenue.”

He’s saying the capex boom isn’t stopping.  Because with each unit of compute added into the world, it’s generating tokens, and tokens are generating revenue.  And revenue equals more capex. It’s proving productive for the consumer, and profitable for the datacenters.

It’s a self-reinforcing loop.

And this new phase of “always-on” AI agents, means more and more economic output, higher productivity, and (related) higher potential economic growth.

That brings us back to Elon’s thesis — his view that humanoid robots and autonomous AI will ultimately remove any meaningful limit on the size of the economy.

The logic is simple: if the cost of labor approaches zero, economic output has no ceiling.

We’ve talked about the Trump administration’s desire for change in Europe.  It has everything to do with countering China’s influence, and restoring U.S./European alignment.

Meanwhile, the Europeans have doubled down on failed policies.  So has Canada.  And Canada, the European Union and Australia have, in recent weeks, talked up the idea of a new “global order” where they could combine forces to become a “third pole,” to match U.S. and China powers.

They’re missing a very important consideration: The world is already in a two horse race for AI supremacy (and they aren’t one of the horses).

As the CTO for Palantir has said, “the AI race is a winner takes all.”

The winner will set standards, attract talent, and determine what technology gets embedded into governments and critical infrastructure.  And that will mean the difference between AI that serves humanity, or AI that controls humanity (serving the interest of the Chinese Communist Party).

As you can see in this chart, the race is very, very tight.  

And if we listen to the guy that runs the company that’s in the lead (Dario Amodei at Anthropic), this growth curve for model intelligence is about to go vertical — the AI is good enough to improve itself (and direct machines) such that progress starts going vertical. 

This is why, at the AI Action Summit in Paris last year, JD Vance rejected the EU’s “global” AI agreement — because it would slow the innovation and hand to global leadership torch to China.  Instead, he said partner with the U.S. or “chain your nation to an authoritarian master” (i.e. China).

Last week, we talked about Europe’s scramble to convince the world it can credibly pursue “strategic autonomy” (in the words of Christine Lagarde, ECB President) — self-funding their defense, energy and AI buildout, without restarting the European sovereign debt crisis of 15-years ago.

Beyond Trump’s use of tariff and military security bargains to influence change in Europe, there’s a more immediate and combustible bargaining chip coming in May — the U.S. financial security chip.  

Under a new Trump-appointed Fed Chair, if implicit U.S. financial support (like access to dollar liquidity) becomes “conditional” for Europe, then they have a problem.  It would quickly expose the liquidity and then solvency vulnerabilities.

That’s why European leaders are pre-emptively running a confidence management campaign, promoting big plans, aggressive timelines and promises.

But why put that pressure on a critically important Western ally?  

As we’ve discussed over the past year, the Trump administration’s desire for change in Europe has everything to do with countering China’s influence, and restoring U.S./European alignment.

Remember, we’ve looked at this Pew Survey several times over the past year.  As you can see, China has gained significant influence over Europe, and largely stemming from its role in bailouts, following the sovereign debt crisis in Europe more than a decade ago.  

This leads us to this Marco Rubio’s speech this past weekend at the Munich Security Conference.

Marco didn’t hold any punches.  He formalized the move from “implicit” to “conditional” support — by calling out Europe’s “foolish” policies on climate, migration, deindustrialization as anti-sovereignty and self-harming.

He reset the alliance on the basis of “reciprocity and seriousness.”

He evoked the history of “godless communist revolutions” and terminal decline of the 1940s Europe.

And he said the U.S. seeks “renewal and restoration” of Western civilization.  This was an invitation to do it together — “join us.” 

Otherwise, the message has become clear that Europe is not entitled to automatic U.S. backstops (Trump’s leverage).

And as JD Vance said last year at the AI Action Summit in Paris, just days into the Trump administration’s term, partner with the U.S. or “chain your nation to an authoritarian master” (i.e. China).  

We’ve talked about Europe this week. 

EU leaders met yesterday with one important job: convince the world that Europe can fund itself, and that its politics can still execute, before markets decide to test the system.

That said, they admitted they have to move fast. 

Why?  The clock is ticking.  Regime change at the Fed will be here officially in three months, and as we’ve discussed in my daily notes, it will create significant vulnerabilities for Europe. 

With that in mind, the “plan” is “One Europe, One market” (consolidating borrowing power, and turning Europe’s savings into domestic investment).  And they set a June deadline, which is public acknowledgement of the urgency. 

But they’re already setting expectations for a fracture.  They’re laying the groundwork for a failure of unanimity. 

The European Commission President, Von der Leyen said if all 27 can’t agree (small chance that they would), they are prepared to move ahead with a smaller group — group of 9.  A coalition of the willing. 

This will be another moment when the structural flaws of the European Union and the European Monetary Union get exposed.  One currency, one central bank — but many fiscal authorities and fragmented decision-making.

The euro area’s stability ultimately depends on (the market’s) confidence that member countries can work together and execute under stress.

For now, markets are still complacent.  European bank stocks have actually outperformed U.S. banks in this cycle (chart below).  If the market starts to become unconvinced Europe can “muddle-through” we’ll see it in this chart (i.e. divergence).   

Interestingly, today (the first trading session after this meeting), we had divergence — European banks down 3.2%, U.S. banks were up 0.5%.

And the divergence showed up in the real stress gauge: spreads.

Italian yields rose on the day while German yields fell — a widening of the risk premium.

Italy remains an X-factor in the broader fiscal integration of Europe. If Meloni ends up among the unwilling, then the market may start to consider the odds that the integration effort could turn into disintegration of Europe. 

We’ve talked about the pressure the incoming Trump-appointed Fed Chair is putting on Europe.

In about three months, the 15-year era of global central bank coordination — where the Fed implicitly backstopped the world via unlimited access to U.S. dollar liquidity — is likely over.

Access to dollars becomes Trump’s leverage to influence change in Europe. 

And without friendly Fed coordination, Europe becomes exposed to another sovereign debt crisis (an existential threat), as they are facing trillions of euros in new spending commitments.

So, the EU leaders are scrambling.  They are meeting tomorrow to game plan a way to self-fund their big spending needs (defense, AI, energy), and shore up confidence in liquidity access.

To have a chance at making it all work, they need “unification.” 

They need to finance on the full strength of Europe (a eurobond).  They need fiscal union.

And they will be consulting two former Italian Prime Ministers that have mapped this out — Mario Draghi (also former head of the ECB) and Enrico Letta. 

Both have been vocal in calling for common EU debt and integrated capital markets as necessary for survival of the euro.

That will be the pitch tomorrow.  But they have a problem.  They need unanimous vote, by law.  

That’s where tomorrow gets interesting. 

Draghi saved the euro in 2012, as ECB President, by ripping up the rulebook.

The Maastricht Treaty said “no bailouts.”  He found a loophole and did it anyway.

In the current situation, under unanimity rules, a single country can veto legislation — in this case, the idea of a common debt issuance. 

But Draghi and Letta have a work around plan (ripping up the rulebook again), to require only a “qualified majority.“

This could mean integration in Europe is coming, or disintegration. 

Last week we talked about the Google and Amazon earnings.

The companies building AI computing supply still can’t build fast enough to meet demand.  That was signal buried under the noise of a Bitcoin decline and pontifications about software obsolescence.

Today, we got another clear signal.

Taiwan Semi reported an explosion in January sales — typically a post-holiday lull month.  They reported 37% year-over-year growth.  But that wasn’t the jaw-dropper — it was the 20% jump in revenue from the prior month (month-over-month growth)!

At the same time, the board gave a greenlight to a $45 billion capex plan — to build out capacity “based on market demand forecasts.”

Remember, Nvidia’s growth over the past two years has only been constrained by TSM’s ability to fulfill Nvidia’s demand for chips.  TSM had hit a capacity wall.  The demand was there, but the manufacturing capacity wasn’t.

That capacity has since been expanded, and will be further expanded with today’s capex plan announcement.

This 19% monthly growth for TSM looks like the “Nvidia moment“ for chips. 

Remember, Nvidia shocked the world in May of 2023, with a huge quarter.  And then Jensen Huang told us the data center demand was so steep, that the next quarter revenues would jump by more than 50% (from $7 billion to $11 billion).

And he said the hyper-growth would continue for the foreseeable future.  So, May 2023 was “the moment” the world realized that AI was about to reinvent computing.

With that in mind, just a couple of weeks ago, we may have seen the “Nvidia moment” for storage — the moment the world realized AI demand for data storage is endless.

Sandisk reported $3 billion in revenue (up 31% from the prior quarter) and then guided for $4.4-$4.8 billion next quarter(!) — a 50% sequential leap in just 90 days.

And now we may have “the moment” for the world’s most advanced AI chips.  And this is the proxy for the health of the entire hardware ecosystem.  It’s on fire.