Nvidia Beat by $4 Billion. Guided $6 Billion Above Consensus. The Stock Rose 4%. After Two Weeks of Beat-and-Sell, That Is the Signal.
★ ELITE MARKET POINT
Market Intelligence That Moves With You
THE VERDICT
Nvidia Beat by $4 Billion. Guided $6 Billion Above Consensus. The Stock Rose 4%. After Two Weeks of Beat-and-Sell, That Is the Signal.
Revenue hit $96.2 billion. Earnings doubled. Q3 guidance topped $108 billion. Jensen Huang said demand is accelerating. The market held the beat for the first time in two weeks. The old playbook died. A new one just started. Jackson Hole is Friday.
Thursday, August 27, 2026 • NVDA: $218.77 (AH +4%) • Core PCE: 3.3% • 30-Year Yield: 5.22% • Jackson Hole: Friday
Key Idea
Nvidia reported $96.2 billion in revenue, crushing the $92.2 billion consensus by $4 billion. Adjusted earnings hit $2.22, doubling from a year ago. Data center revenue reached $89 billion, above the $85.7 billion estimate. Gross margin held at 75%. Q3 guidance came in at $105.8 billion to $110.1 billion, roughly $6 billion above the $103.9 billion consensus at the midpoint. Jensen Huang said six words that matter more than any number: “AI has reached its inflection point.” Then he added: “Compute is revenue. And demand is accelerating.” The stock rose 4% after hours. Not 12%. Not 15%. Four percent. On any other day, that would be a disappointment. After two weeks of every company that beat selling off, the market holding a Nvidia beat is the most important signal of the year.
THE NUMBERS THAT BROKE THE PATTERN
Revenue: $96.2 billion. Up 106% year over year. Up 18% quarter over quarter. Last year, same quarter, revenue was $46.7 billion. Nvidia doubled its revenue in twelve months. No company of this size has ever done that.
Data center: $89 billion. That is where the AI money lives. Hyperscaler revenue from Amazon, Google, Microsoft, and Meta drove the number. Enterprise AI customers grew. Sovereign AI demand, governments building their own AI infrastructure, accelerated. Nvidia announced a major expansion of its strategic collaboration with AWS on the call.
Earnings per share: $2.22 adjusted, versus $2.09 consensus. Revenue beat by $4 billion. EPS beat by 6.2%. Gross margin held at 75% despite the Blackwell production ramp. That was the number we told you to watch. Blackwell is a larger, more complex chip. Manufacturing yields at TSMC were a question mark. The answer: margins held. The bears lost their best argument.
Q3 guidance: $105.8 billion to $110.1 billion. The midpoint is $108 billion. Consensus was $103.9 billion. That is a $4 billion beat on the forward guide. In a market that stopped paying for backward-looking beats, Nvidia delivered a forward-looking one. That is why the stock went up instead of down. AMAT, SNDK, TGT, and WMT all beat on earnings and missed or guided weakly. Nvidia beat on earnings and guided aggressively. The forward guide broke the pattern.
Huang went further. He said Nvidia expects revenue to grow roughly 70% in fiscal 2028. CPU revenue will more than double. Supply will remain a bottleneck through the end of FY28. That last detail is the most bullish sentence in the entire report. When demand outstrips supply for two more years, pricing power holds. When pricing power holds, margins hold. When margins hold, the stock holds.
Data center: $89 billion. That is where the AI money lives. Hyperscaler revenue from Amazon, Google, Microsoft, and Meta drove the number. Enterprise AI customers grew. Sovereign AI demand, governments building their own AI infrastructure, accelerated. Nvidia announced a major expansion of its strategic collaboration with AWS on the call.
Earnings per share: $2.22 adjusted, versus $2.09 consensus. Revenue beat by $4 billion. EPS beat by 6.2%. Gross margin held at 75% despite the Blackwell production ramp. That was the number we told you to watch. Blackwell is a larger, more complex chip. Manufacturing yields at TSMC were a question mark. The answer: margins held. The bears lost their best argument.
Q3 guidance: $105.8 billion to $110.1 billion. The midpoint is $108 billion. Consensus was $103.9 billion. That is a $4 billion beat on the forward guide. In a market that stopped paying for backward-looking beats, Nvidia delivered a forward-looking one. That is why the stock went up instead of down. AMAT, SNDK, TGT, and WMT all beat on earnings and missed or guided weakly. Nvidia beat on earnings and guided aggressively. The forward guide broke the pattern.
Huang went further. He said Nvidia expects revenue to grow roughly 70% in fiscal 2028. CPU revenue will more than double. Supply will remain a bottleneck through the end of FY28. That last detail is the most bullish sentence in the entire report. When demand outstrips supply for two more years, pricing power holds. When pricing power holds, margins hold. When margins hold, the stock holds.
PCE CAME IN HOT. THE MARKET SHRUGGED.
Before Nvidia reported, the Bureau of Economic Analysis released July PCE data at 8:30 AM. Headline PCE: 3.7% year over year, one tenth above expectations. Core PCE: 3.3% year over year, matching the forecast. Monthly core: 0.2%, in line.
The headline miss was driven by energy. Gasoline and energy goods spending fell $14 billion in July, the largest single-category decline in the report. But services prices rose 0.3%, pushed by a 1.2% increase in financial services and insurance. Housing rose 0.3%. The sticky categories stayed sticky.
Here is the detail that matters for Jackson Hole. Core PCE has been 3.3% for four consecutive months. April: 3.3%. May: 3.4%. June: 3.3%. July: 3.3%. No progress. The Fed’s target is 2.0%. Core inflation is stuck 130 basis points above target and has not budged since spring. That is the number Warsh will have in his hand when he speaks Friday morning.
September hike odds dropped to 36% after the report. The market interpreted core PCE matching expectations as good enough. But “good enough” at 3.3% is not progress. It is a plateau. And a plateau at 3.3% with the 30-year yield at 5.22% means Bessent’s two-front war continues. The bond market is not satisfied. The Fed has not declared victory. And Warsh has a blank piece of paper.
The headline miss was driven by energy. Gasoline and energy goods spending fell $14 billion in July, the largest single-category decline in the report. But services prices rose 0.3%, pushed by a 1.2% increase in financial services and insurance. Housing rose 0.3%. The sticky categories stayed sticky.
Here is the detail that matters for Jackson Hole. Core PCE has been 3.3% for four consecutive months. April: 3.3%. May: 3.4%. June: 3.3%. July: 3.3%. No progress. The Fed’s target is 2.0%. Core inflation is stuck 130 basis points above target and has not budged since spring. That is the number Warsh will have in his hand when he speaks Friday morning.
September hike odds dropped to 36% after the report. The market interpreted core PCE matching expectations as good enough. But “good enough” at 3.3% is not progress. It is a plateau. And a plateau at 3.3% with the 30-year yield at 5.22% means Bessent’s two-front war continues. The bond market is not satisfied. The Fed has not declared victory. And Warsh has a blank piece of paper.
JACKSON HOLE TOMORROW: WHAT WARSH KNOWS
Fed Chair Kevin Warsh speaks Friday morning at the Jackson Hole Economic Policy Symposium. He has given zero forward guidance since taking office. His speech is, in his own words, “a blank piece of paper.”
Here is what he knows as he writes it. Core PCE has been flat at 3.3% for four months. The 30-year yield hit a 19-year high two weeks ago and Bessent had to intervene three times to bring it down. Three FOMC members dissented for a hike in July. Several more were sympathetic. Oil has fallen 5% this week but remains $20 above pre-war levels. Nvidia just reported the largest revenue quarter in semiconductor history and said demand is accelerating through 2028.
A Bank of America survey shows 69% of fund managers expect a neutral tone. Neutral is priced. The surprise is what moves markets.
If Warsh mentions the 30-year yield, even once, traders will interpret it as tacit support for Bessent’s buyback intervention. Yields ease. The S&P pushes higher. If he stays focused on inflation and signals that September is live, the 30-year retests 5.34%. If he pivots to structural concerns about AI capital spending and financial stability, that is a new variable nobody has priced.
The speech starts Friday morning. PCE is in his hand. Nvidia’s quarter is on his desk. The two-speed economy we identified twelve days ago is now the consensus framework. Warsh decides whether the Fed validates it or fights it.
Here is what he knows as he writes it. Core PCE has been flat at 3.3% for four months. The 30-year yield hit a 19-year high two weeks ago and Bessent had to intervene three times to bring it down. Three FOMC members dissented for a hike in July. Several more were sympathetic. Oil has fallen 5% this week but remains $20 above pre-war levels. Nvidia just reported the largest revenue quarter in semiconductor history and said demand is accelerating through 2028.
A Bank of America survey shows 69% of fund managers expect a neutral tone. Neutral is priced. The surprise is what moves markets.
If Warsh mentions the 30-year yield, even once, traders will interpret it as tacit support for Bessent’s buyback intervention. Yields ease. The S&P pushes higher. If he stays focused on inflation and signals that September is live, the 30-year retests 5.34%. If he pivots to structural concerns about AI capital spending and financial stability, that is a new variable nobody has priced.
The speech starts Friday morning. PCE is in his hand. Nvidia’s quarter is on his desk. The two-speed economy we identified twelve days ago is now the consensus framework. Warsh decides whether the Fed validates it or fights it.
THREE SCENARIOS FOR THE OPEN
Base
Nvidia opens 4–6% higher near $220–$225. SOX rallies 3–4%. The S&P pushes above 7,750. The AI trade gets its reprieve. But the move is measured, not euphoric. At 25 times forward earnings with a $108 billion Q3 guide, the stock is fairly valued, not cheap. Jackson Hole is 24 hours away. Nobody wants to be aggressively long going into Warsh’s speech. You hold NVDA if you own it. You do not chase above $225 until Warsh clears.
Upside
Nvidia opens 8–12% higher above $230. The $108 billion guide triggers a wave of analyst upgrades. SOX rallies 5–7%. The S&P pushes above 7,850 and approaches its all-time high. Warsh speaks neutral or dovish Friday. September hike odds collapse below 25%. The 30-year yield drops below 5.10%. The two-speed economy has its engine and the Fed is not going to shut it off. Jay Woods’s path to S&P 8,000 opens. If you own NVDA, SMH, and XLK, you hold through the weekend.
Risk
Nvidia opens higher but fades through the session as traders sell into strength ahead of Jackson Hole. The after-hours gain evaporates by noon. The pattern from August 2024 and 2025 repeats: beat, pop, fade, sell. Warsh speaks hawkish Friday and signals September is live. The 30-year yield retests 5.30%. Core PCE at 3.3% for four straight months gives the hawks ammunition. The S&P gives back the Thursday gains. Defense and gold catch the bid. The verdict from Nvidia was positive but the verdict from the Fed overrides it. If you own NVDA, you hold through the fade but you do not add until Warsh clears.
Our View
Twelve editions. Twelve days. Every call landed. And the verdict is in.
On August 17 we told you the market and the consumer were pricing two different economies. On August 18 we told you the bond market would win that argument. On August 19 we told you the rotation from momentum to durability had begun. On August 20 we told you Bessent blinked. On August 21 we told you Walmart confirmed the consumer is rationing. On August 22 we told you the economy was booming and the consumer was cracking at the same time. On August 23 we told you the old playbook was dead. On August 24 we told you Nvidia was the test. On August 25 we told you Bessent was fighting a two-front war. Yesterday we told you the market cleared the runway. Today we have the result.
Nvidia held.
That is the verdict. Revenue doubled. Guidance crushed consensus by $6 billion. Margins held. Supply is bottlenecked through 2028. And the stock went up. Not violently. Not euphorically. Steadily. Four percent. After AMAT fell 7.5% on a record quarter. After SNDK fell 9% on 372% growth. After WMT fell 9% despite beating every line. Nvidia beat bigger than all of them and the stock went up.
The pattern broke. The beat-and-sell trade that defined the past two weeks ended on the one stock that could end it. The AI trade is alive. The growth engine of the two-speed economy just proved it has fuel through 2028.
But the other speed has not changed. Core PCE is stuck at 3.3%. The 30-year yield is 5.22%. Bessent has intervened three times in ten days. Walmart’s CFO said consumers are rationing. Michigan sentiment is at the 2nd percentile. Gas is $4.08. The boardroom is booming. The kitchen table is not.
Tomorrow Warsh speaks. He has the Nvidia quarter on his desk and 3.3% core PCE in his hand. He will decide whether the Fed validates the two-speed economy or fights it. That is the last question of these twelve days. We will be here tomorrow with the answer.
On August 17 we told you the market and the consumer were pricing two different economies. On August 18 we told you the bond market would win that argument. On August 19 we told you the rotation from momentum to durability had begun. On August 20 we told you Bessent blinked. On August 21 we told you Walmart confirmed the consumer is rationing. On August 22 we told you the economy was booming and the consumer was cracking at the same time. On August 23 we told you the old playbook was dead. On August 24 we told you Nvidia was the test. On August 25 we told you Bessent was fighting a two-front war. Yesterday we told you the market cleared the runway. Today we have the result.
Nvidia held.
That is the verdict. Revenue doubled. Guidance crushed consensus by $6 billion. Margins held. Supply is bottlenecked through 2028. And the stock went up. Not violently. Not euphorically. Steadily. Four percent. After AMAT fell 7.5% on a record quarter. After SNDK fell 9% on 372% growth. After WMT fell 9% despite beating every line. Nvidia beat bigger than all of them and the stock went up.
The pattern broke. The beat-and-sell trade that defined the past two weeks ended on the one stock that could end it. The AI trade is alive. The growth engine of the two-speed economy just proved it has fuel through 2028.
But the other speed has not changed. Core PCE is stuck at 3.3%. The 30-year yield is 5.22%. Bessent has intervened three times in ten days. Walmart’s CFO said consumers are rationing. Michigan sentiment is at the 2nd percentile. Gas is $4.08. The boardroom is booming. The kitchen table is not.
Tomorrow Warsh speaks. He has the Nvidia quarter on his desk and 3.3% core PCE in his hand. He will decide whether the Fed validates the two-speed economy or fights it. That is the last question of these twelve days. We will be here tomorrow with the answer.