Nvidia Answered the Numerator. Two Days Later the Fed Chair Raised the Denominator.

The Week Delivered Both of Its Answers

On Monday this column posited that two unrelated events, forty eight hours apart, would define the rest of the year.

Nvidia would answer the question of whether the artificial intelligence buildout is real.

The new Fed chair would answer the question of what price the buildout would be earned at.

Both have now happened, and both answers were yes.

Nvidia's was emphatic.

Kevin Warsh's, delivered this morning from Jackson Hole, was that the price of money is more likely to go up than down.

They are not contradictory conclusions, but the dual sides of a valuation worksheet, and the market spent today realizing the second matters far more than the first.

Wednesday Night: 96.2 Billion Dollars

Nvidia reported revenue of 96.2 billion dollars for the quarter ended July 26, up 18 percent sequentially and 106 percent year over year.

Focus on that second figure.

The company at least doubled its revenue in twelve months.

Consider that a company of this size did not merely scale its operations from a start up to a major player, but added 50 billion dollars to its quarterly revenues against the same period last year.

Revenue from data centers was 89.0 billion dollars, up 117 percent year over year and 18 percent sequentially, driven by the company's Blackwell Ultra graphics processors.

This product category makes up 92 percent of Nvidia's total sales.

Hyperscale revenue, sales to the few dominant clouds, more than doubled in the past year, while edge computing, an important but smaller segment of the company's business, was 7.2 billion dollars, up 27 percent year over year.

GAAP EPS was 2.46 dollars, and adjusted EPS was 2.22 dollars, compared to 1.05 dollars a year ago.

Gross margins were 75.0 percent.

The company returned 26.0 billion dollars to shareholders in the form of buybacks and dividends during the course of the quarter, with 99.0 billion dollars in shares repurchases remaining authorized, and it has scheduled a quarterly dividend of 0.25 dollars per share, payable October 1.

The Guide Was the Real News

None of this moved the stock, because the results were those of a company that had a great quarter.

The market had already priced in a strong result.

What mattered was what came next.

Nvidia guided for 108.0 billion dollars in revenue in the current quarter, give or take 2 percent, which is another 12 billion dollars on top of the 96 billion last quarter, and another 12 percent sequential increase on a base that just grew 18 percent.

Jensen Huang then spoke to fiscal 2028, estimating revenue growth of roughly 70 percent, on a fiscal 2027 base that is already tracking toward roughly 400 billion dollars.

The consensus among analysts followed by LSEG had been 44 percent.

A company guiding 26 percentage points above the analyst median for a year not yet begun says something about the world, and not merely about itself.

Mr. Huang said as much on the call, stating that "AI has reached its inflection point. It's doing useful work. Its tokens are productive and profitable. Now, compute is revenue."

What 75 Percent Gross Margin Tells You

Gross margin is revenue minus cost of goods sold, divided by revenue.

It is the single most important number in a public company's earnings deck, because it is a measure of pricing power.

A 75 percent gross margin is extraordinary.

It suggests that for every dollar of sales, Nvidia retains 75 cents before paying a penny to its employees or suppliers or landlords.

Few industries have such pricing power.

An industrial manufacturer that operates at 30 percent gross margin, or a software company at 80 percent, would kill to have such a business model.

Nvidia is a hardware company with the margins of a software company, and that suggests that customers have limited options, and are less sensitive to price.

Because Nvidia's customers are hyperscalers, companies that are building out cloud infrastructure at an astonishing rate, and competing intensely to capture market share, they have little time to shop around and every incentive to pay a premium for capacity.

The real value of a 75 percent gross margin is that it is an invitation to competition.

It creates an opening for other firms, large and small, to participate in the market.

That is why hyperscalers are designing their own chips, and it is not a threat to Nvidia's margins.

But it is also why those margins cannot last.

While they are a short term advantage to the company, they are a long term subsidy to the entire industry, one that disappears as soon as supply catches up with demand.

Thursday Belonged to the Buildout

Markets took the news in stride.

Nvidia's shares rose 7.4 percent in premarket trading on Thursday, and closed the day up 8.7 percent.

The broader market followed suit.

Broadcom was up 4.5 percent, while Intel gained 4 percent in after hours trading.

In Asia, SK Hynix added 2 percent, while in Europe, the Stoxx technology index gained 1.8 percent.

The reaction was important, because it was not specific to Nvidia.

One company's stock result is a company scorecard, but when one company's result improves memory manufacturers in Korea and equipment makers in Europe, it is an industry re-rating.

That is what happened to the market this week.

For one day, at least, the question facing the market that this column posed on Monday had an affirmative answer.

The buildout was real, the cash was real, and companies at the leading edge of the transition were finding ways to convert both to earnings, rather than merely projecting them.

And Then There Was Friday

Kevin Warsh, who delivered his first Jackson Hole speech this morning as the new Federal Reserve chair, spoke to a market that was not prepared for him to say much of anything at all.

The usual custom in Jackson Hole is that the chair speaks, and that markets listen.

They did not have to listen hard.

Mr. Warsh declined to give them much to work with, saying instead that he had no intention of providing forward guidance.

What he did was speak to prices, and to the appropriate communication strategy for discussing them, and both were unexpectedly hawkish.

On prices he was blunt. "Inflation is running above our 2 percent target," he said. "So the Fed's predominant focus right now should be on prices."

He noted that the PCE price index was running at 3.7 percent over the past twelve months and 4.1 percent over the most recent six.

Recall that the most recent inflation number, which many interpreted as a sign that prices were cooling, was a six month number, and that it was in fact accelerating.

On the benign interpretation of the summer numbers, he was equally unmoved.

"Those readings do not tell me that underlying trends have meaningfully improved," he said.

Hard Pressed to Call It Restrictive

The sentence that caused the market to wobble was this one: "I would be hard pressed to describe broad financial conditions as restrictive."

When the speaker is the Federal Reserve chair, and the statement concerns financial conditions, it is worth listening carefully.

By stating that he was hard pressed to describe them as restrictive, Mr. Warsh was announcing that they were not.

The phrase "financial conditions" is broad, but it typically refers to the general ease of credit, and Mr. Warsh gave several reasons why current conditions were easy.

He noted that credit spreads, the difference between investment grade bonds and Treasuries, were near a generational low, and that corporate lending standards were also easy.

Business investment was robust, and he concluded that a policy rate of between 3.50 and 3.75 percent was not, in fact, restrictive.

In other words, the economy had adapted to higher interest rates, and so had corporate America.

If six month inflation was running at 4.1 percent, while the federal funds rate was only 3.5 percent and conditions were not restrictive, there was only one reason for that.

The End of Forward Guidance

The second half of the speech was institutional, touching on the end of the forward guidance doctrine, and it will live on in the market long after the first half has been forgotten.

Mr. Warsh suggested that the Federal Reserve should cease issuing forward guidance, the formal statements about future policy moves, suggesting that the practice has "overstayed its welcome."

His reasoning was based on a hall of mirrors theory in which the Federal Reserve and the market look at each other, neither willing to admit to a change in policy, and neither able to react to one when it comes.

"We should not indulge a regime in which market participants are looking primarily to the Fed for their next trade."

He also announced the formation of five internal task forces to study monetary policy, including one dedicated to the economic effects of artificial intelligence.

The second part of this speech is far more important than it appears, because it touches directly on the issues discussed in this column.

The announcement that the Federal Reserves studies the economic effects of artificial intelligence suggests that it has recognized the buildout as more than a technological or sector specific phenomenon, and one that touches deeply on labor markets and the neutral rate of interest.

How the Market Actually Took It

The market's headlines are always small.

The S and P 500 finished the day down 0.25 percent at 7,711, while the Nasdaq fell 0.5 percent to 26,402.

The Dow was little changed, down just 10 points at 53,559.

The moves in the headline indices were small, but moves underneath were not.

In the futures market, the implied probability of a rate hike at the September meeting jumped from just 35 percent to roughly 57 percent, meaning that the market now expects the Federal Reserve to tighten monetary policy, and soon.

In the Treasury market, the yield on the two year Treasury note, which reflects policy expectations most directly, jumped over 6 basis points to 4.298 percent, a basis point being one hundredth of a percentage point.

At the same time, yields at the long end of the curve declined.

The combination of higher yields at the short end and lower yields at the long end is known as a flattening of the curve, and it suggests that the market expects the Federal Reserve to raise interest rates soon, and to leave them higher for an extended period of time.

It is not a vote of confidence.

It is a recognition that tightening will come, and that it will hurt.

Why Smaller Caps Were Down Five Times the S And P

The Russell 2000, a broad based index of smaller American companies, was down 1.30 percent to 2,975.13, more than five times the decline of the S And P 500.

The reason has to do with balance sheets, and what the change in expectations means for them.

Large companies typically finance themselves with long term bonds, issued at a fixed rate, while small companies tend to rely on bank financing.

The two structures have very different risks.

When the yield on a two year note jumps, it increases the financing costs of small companies, but the interest costs of large companies tend to remain the same.

The second reason is that smaller companies tend to carry more debt on their balance sheets, and have less cash than their larger counterparts.

A hike in short term rates affects their expenses the most.

A rise in rates next month would affect the expenses of small caps most immediately.

Their costs would rise, their cash would be reduced, and their future earnings power would be discounted.

The Russell 2000 told the market what a hawkish speech from the Fed actually meant.

There was a second lesson, and it was that the entire week's worth of speeches and earnings releases were linked by a common theme.

A company's value is expressed as a function of its future cash flows, discounted by a rate that reflects the cost of capital.

On Wednesday night, Nvidia increased the numerator, and announced its intention to do so again.

On Friday morning, the Federal Reserve increased the denominator.

The market spent today deciding that both actions were correct, and that the second would apply to everyone, while the first only applied to a few.

The Bottom Line

The AI trade got the confirmation it needed.

Nvidia grew revenue 106 percent, held 75 percent gross margins, guided for 108 billion dollars next quarter, and told investors to expect 70 percent year over year growth in a fiscal year that has not yet arrived.

These are not projections about a future technology, but statements about current cash flows and existing revenues.

The cost of money got a hike as well, in the other direction.

The Fed chair said that conditions were not restrictive, that six month inflation was running at 4.1 percent, that the improvement in the summer was an illusion, and that markets should stop looking for forward guidance and start realizing that hikes were coming.

Odds for a rate increase in September jumped from just 35 percent to a clear majority in a single morning.

Investors have spent the last two years discussing the AI trade and the rate trade as if they were separate.

This week they were both priced on the same sheet, and the companies that use floating rate debt to finance themselves learned the cost of that structure and liquidity risk the fastest.