Let’s talk about Nvidia earnings …

As we’ve discussed for much of the past two years, data center revenue has been telling a very clear story:  There has been a manufacturing capacity issue.

We’ve looked at this drumbeat of about $4 billion in revenue growth, quarter-by-quarter, for much of the past two years (the red line) …

This quarterly growth ceiling led to a trend of decelerating year-over-year growth, which called into question valuation.

Is Nvidia worth 50+ times earnings if growth continues to slide?

Then there was Q2.

In the August report, not only was that $4 billion drumbeat disrupted — cut in half, to $2 billion — the “compute” component of data center revenue (which has been the growth story for Nvidia) actually contracted.

Certainly it’s not worth 50x earnings after that, right?

However, we went into today’s earnings with that premium valuation proving very durable.

And today, it was validated.

As we suspected, based on some very clear clues left by Jensen a few weeks ago, growth is back in data center.

They did $51 billion in data center revenue in Q3.  It was the hottest quarter-over-quarter growth in seven quarters.  And it was led by compute — the compute component grew by $10 billion, up 27% on the quarter.

So, the market should like the validation from this report.  Growth in GPU deliveries/compute has returned.

And the focus now turns to the other big number that was revealed by Jensen a few weeks ago. In the earnings call today, management affirmed $500 billion of revenue “visibility” through the end of calendar year 2026.

That’s half a trillion dollars of Nvidia’s most advanced chips/systems to fulfill over the next five quarters– or $100 billion a quarter.

Can they fulfill it?

They say the parts are lined up, Blackwell is already shipping in volume, new U.S. wafer production is running, and partners are ready to build.  So, they believe they can deliver.

As we discussed in my last note, with the government now reopened, markets are back to trading economic data. And the risks are skewed toward negative surprises, given that the White House has said it expects the shutdown to shave 100 to 150 basis points from GDP.

We’ll get the September jobs report on Thursday morning. It will reflect conditions from two months ago, but should show a continuation of the four-month trend of very weak job growth.

On that note, we heard from the Fed’s Chris Waller today on his view of labor market conditions. Remember, Waller is one of three on the list to become the next Fed Chair. He gave a prepared speech today titled “The Case for Continuing Rate Cuts.”

Waller sees “underlying” inflation (excluding estimated one-off tariff effects) as near the Fed’s 2% target. That means the Fed funds rate is 75 to 100 basis points too restrictive, given where they see the neutral rate now (3%, from the most recent Summary of Economic Projections).

With that, he says he’s worried that “restrictive monetary policy is weighing on the economy.”

He sees a slower economy in the second half, even before accounting for the government shutdown. And he thinks the labor market has swung from a supply problem to a demand problem (i.e. weaker demand for labor).

So Waller makes the case for another quarter-point cut in December.

Meanwhile, as we discussed last week, the interest-rate market has been pricing out the probability of a December cut.

That weighs on stocks, and is providing fuel for the setup we flagged in late October for a technical correction.

Here’s the latest chart on the S&P …

Remember, as we’ve discussed over the past few weeks, the October 29 peak in the S&P forward P/E was 23. The last time we saw that valuation was September 2, 2020, which was followed by a 10% price correction.

This time around, the technicals are also lined up for a price correction.

From the chart above, we have a bearish technical reversal signal (circled), and a breakdown of the trend from the tariff-shock lows. Add to this: the VIX (the cost of downside stock-market insurance) is back above 20 — a level that tends to come with lower stock prices.

A 10% correction in the S&P (from October highs) would take us back to the August levels of ~6,250.

Another chart to watch is Bitcoin, particularly given that it now sits on corporate balance sheets and is being used as collateral …

As the overlay above shows, Bitcoin and S&P futures have traded in a tight relationship in this cycle. And Bitcoin has already had a sharp and deep drawdown. 

It matters for stocks because Bitcoin is no longer just retail speculation, corporates and financial firms hold it on balance sheet, and it is increasingly pledged against credit lines, structured products and derivatives.

So, when the price falls this quickly, lenders call for more collateral, the easiest thing to sell to raise cash is liquid equities.

In my last note, we talked about some pro-liquidity fiscal and monetary news late last week that was favorable for asset prices.

The FOMC Vice Chair, John Williams, signaled to markets that the Fed would be prepared to expand the balance sheet again in the near future, to avert any 2019-like liquidity shock (only days after announcing the end of balance sheet contraction).  Add to that the Fed’s Miran had more dovish comments Friday, and Trump floated a plan for $2,000 government checks over the weekend.

These types of comments tend to be good for financial asset prices.  With that, stocks had a nice bounce from the Friday morning lows.

But the bigger move has come in hard asset prices.  Gold is up 6% from Friday’s lows.  Palladium is up 9%, and silver is up 12%.

Remember, just two weeks ago, we looked at this technical break down in the price of gold after trading as high as $4,358 in mid-October (record levels).

And we followed this parabolic move in gold over the prior six weeks (a nearly 30% rise in two months). 

Importantly, the timeline of that price surge mapped closely to the EU’s plan to send €140 billion of frozen Russian assets to Ukraine.

Not coincidentally, the top in gold aligned with reports that the Trump administration would NOT support the EU’s plan to use (confiscate) Russian assets.  Trump later reiterated that stance. 

That reduces the “confiscation/distrust premium” being priced into gold — the price fell 10% in about a week. 

And then on this past Friday, we get this …

And we get this price response …

The “confiscation/distrust premium” in gold ramps, again.

As we discussed, given the technical setup and the historically high broad market valuation, things were setup for a technical correction.

But we’ve since had a few pro-liquidity shots across the bow (pro-liquidity = pro-asset prices). 

Last Friday morning the Fed Vice Chair, John Williams, was in Europe doing a keynote at a European Central Bank conference on money markets. 

Just a week after the Fed ended quantitative tightening, he signaled to markets that the Fed would be increasing the balance sheet again in the near future, as part of its (don’t call it QE) “reserve management” framework.

Later on Friday, Fed governor Miran (Trump’s guy) gave a speech on the impact of stablecoins on Treasury demand (higher) and the neutral interest rate (lower). 

This coming structural increase in the demand for short-term Treasurys (estimated in the $1-$3 trillion range) behaves like QE — as Miran said, “may increase the supply of loanable funds in the U.S. economy.”

Finally, on the fiscal side, Trump floated the idea of $2,000 “tariff dividend” checks today. 

This seems like it was directed at the Supreme Court, to apply additional pressure on their decision about the legality of the tariff policies.  Trump is effectively promising the American people the money that the Supreme Court might rule needs to be returned to sender.  Nonetheless, the market is now contemplating the impact of government checks on asset prices and consumption.

With all of this, the trend break in stocks was short-lived, for now.

More on the Fed Vice Chair’s speech …

This speech was about averting another 2019 shock event.

Remember, back in 2018-2019, the Fed spent eighteen months shrinking the balance sheet, draining liquidity from the financial system, and they created a cash crunch (a scramble for dollars in the interbank lending market).

Let’s take a look at a few charts.

Remember, earlier this week we looked at this valuation chart on the S&P 500 (from FactSet) …

This latest peak in the forward P/E was October 29th.  And as we dicussed, a forward P/E above 23 has happened only one other time on this chart — September 2, 2020.

On that day in 2020, stocks started a sharp 10% correction that spanned just 15 days, peak-to-trough.  And that correction came after the 65% rise over the preceding six months.

Given that valuation backdrop, we talked about the set up for a technical correction for stocks.

With that, let’s take a look at the price chart of the S&P 500 …

The high was marked on October 30th, a day that also resulted in a classic technical reversal signal (a bearish outside day – a session with a higher high and lower low than the prior day, closing near the low).  And today, we test this big trendline support which comes in from the “tariff-shock” lows of April.

We have a similar chart in the Nasdaq.

In this case, it’s been a 60% rise over a six month period (echoing the 2020 analogue).  And we get the bearish outside day, at the top, and now the challenge of this big trendline.

The Supreme Court heard two-and-a-half hours of arguments on the legality of Trump-era tariffs today. 

I ran the full transcript through several top AI models. And here’s how the models expect the Court to land …

OpenAI’s ChatGPT-5: Projects a 7–2 split against the government; (International Emergency Economic Powers Act) IEEPA’s “regulate importation” does not authorize revenue-raising tariffs.

Google’s Gemini 2.5: Sees a majority for the challengers, likely 6–3 or 5–4, holding IEEPA doesn’t authorize tariffs.

Anthropic’s Claude Sonnet 4.5: Expects 6–3 or 7–2 against the government; “regulate importation” doesn’t include imposing tariffs to raise revenue.

Perplexity: Anticipates 6–3 or 5–4 against broad presidential tariff authority under IEEPA.  Many Justices expressed skepticism about finding such broad power in the “regulate importation” language of the statute, without explicit Congressional authorization.

xAI’s Grok: Goes 8–1 against the administration, potentially forcing reliance on narrower authorities (like Section 232) and even phased refunds.

Bottom line from the models:  consensus view is a majority vote against Trump tariffs. 

So, would that undo the past 10 months of deal-making or unwind the global realignment achieved under Trump’s use of tariffs as negotiating leverage? 

It shouldn’t. 

Many of the deals are locked in, via bilateral agreements.  Some already fit into this “narrower” authority, where tariffs on things like steel and aluminum are legal under section 232 of the Trade Expansion Act of 1962, as threats to U.S. national security.  And broadly, the expectation is that there will be a migration to other statutes which support the tariff strategy — that just weren’t as convenient and as quick to execute as IEEPA.

With the above in mind, we get a move in Treasury yields today (higher), as markets reconsidered the deficit/revenue path — given potential refunds and a weaker outlook on tariff revenue.

In yesterday’s note, we talked about the concern over “intra-industry” funding among the AI giants.

It’s feeding fears of an AI bubble.

But as we discussed, it looks more like a feature of the AI infrastructure boom, not a bug.  By coordinating capital, compute capacity and AI deployment, the momentum in AI infrastructure development continues, and with that, the probability that America wins the race for global AI leadership increases.

And when the point of overbuilding/overspending is hit, the hyperscalers will dial back capex, which just means their free cash flow goes UP (which is already in the 10s of billions of dollars every quarter).  

They are already cash flow machines.  When AI adoption is widespread (and data centers are humming around the clock) they will generate even more cash.

More broadly, two-thirds of the S&P 500 have reported Q3 earnings, and the numbers are very good: 83% have beaten earnings estimates and 79% have beaten revenue estimates.  And earnings growth is tracking near 11%, which would be four consecutive quarters of double digit earnings growth.

So, earnings are good.  The economy is good.  The Fed is in an easing cycle.  All good for stocks.

That said, we spent some time last month talking about the setup for a technical correction in stocks.  And there were technical and fundamental reasons to believe Nvidia would lead it.  Nvidia did have a 10% correction over nine short days.  Broader stocks had a much shallower decline.

But we still have this chart from FactSet on valuation …

Last week (October 29th), the forward P/E on the S&P 500 was above 23.  As you can see, that’s happened only one other time on this chart — on September 2, 2020. 

What happened on September 2, 2020?   

Stocks started a 10% correction on that day, for a duration of 15 days peak-to-trough.  And that correction came after the 65% rise from the March 2020 lows (the depths of the covid lockdown-induced crash in stocks).

The economy was bouncing back very strongly.  There was war rhetoric around China.  But this was a technical correction.

So, given this valuation data point, are we in the early stages of a technical correction for stocks right now? 

If we look at the October 30th high in the S&P 500, it did come with an outside day (a technical reversal signal).  Stocks are down 3% since.

As we discussed last month, the garden variety 10% corrections of the past two years have resulted in a Fed reaction (either signaling or direct policy action).  In the current case, that would mean the Fed restoring expectations of a December cut.