Nvidia Stock Q2 Earnings Analysis: NVDA Finally Breaks Its Post-Earnings Losing Streak

Nvidia has finally done what it failed to do after four consecutive earnings reports: beat Wall Street’s estimates and make NVDA stock rise. Revenue crossed $96 billion, the next-quarter outlook comfortably cleared expectations and management projected roughly 70% growth for fiscal 2028. 

However, this was not a perfect report. Profit margins are heading lower and cash collection slowed sharply. The market simply decided that Nvidia’s demand outlook was strong enough to outweigh those concerns.

Let’s break down why NVDA stock rose 4.71% after earnings, what Nvidia got right and the new financial risks hiding behind its record AI-chip sales.

Nvidia Q2 FY2027 Earnings Results: Revenue, EPS and Key Numbers

Nvidia reported its fiscal second-quarter results for the three months ended July 26, 2026.

Sources: Nvidia’s earnings release

Nvidia beat the revenue consensus by approximately $4.05 billion or 4.4%. Adjusted EPS exceeded expectations by about 5.7%, while Data Center revenue came in nearly $3.94 billion above consensus.

More importantly, Nvidia cleared every major revenue hurdle discussed in our earnings preview.

Nvidia exceeded its own guidance midpoint by $5.22 billion and even beat the top of its guidance range by $3.40 billion.

This was not a quarter that merely looked good against an easily managed forecast. It also cleared the higher unofficial bar investors had placed on the company.

Why Is Nvidia Stock Rising After Earnings?

NVDA stock closed 1.59% lower at $209.66 on August 26 before the results were announced. It then increased 4.71% to $219.53 in after-hours trading.

The stock initially fell after the earnings release because Nvidia’s gross-margin outlook was weaker than expected. It reversed during the conference call as management provided a much stronger longer-term growth forecast.

Five developments drove the turnaround.

1. Nvidia beat both the official and unofficial earnings bars

Wall Street was already expecting Nvidia to beat its $91 billion guidance. Therefore, a small beat may not have been sufficient.

Instead, Nvidia reported $96.22 billion. This was approximately $4 billion above consensus and around $1 billion to $2 billion higher than the more optimistic $94 billion to $95 billion expectations discussed before the results.

The scale of the beat made it difficult to dismiss the quarter as another predictable Nvidia beat.

2. Q3 guidance was nearly $4 billion above consensus

Nvidia guided for $108 billion of Q3 revenue, plus or minus 2%. This produces a range of $105.84 billion to $110.16 billion.

Wall Street had expected approximately $104.19 billion, according to LSEG estimates reported by Reuters.

The guidance midpoint is:

• Approximately $3.81 billion above consensus
• Around 12.2% higher sequentially
• Approximately 89% higher year on year

Importantly, Nvidia’s guidance assumes no Data Center compute revenue from China. Therefore, the $108 billion forecast does not depend on a recovery in one of Nvidia’s most uncertain markets.

3. Nvidia expects around 70% revenue growth in FY2028

This was probably the most important statement on the earnings call.

CFO Colette Kress said Nvidia expects revenue to grow by approximately 70% in fiscal 2028, despite remaining supply-constrained. FactSet’s earlier consensus had expected growth closer to 45%.

The stock moved higher after this forecast was disclosed, according to Barron’s earnings coverage.

This changed the investor conversation. Before earnings, the question was whether Nvidia could reach quarterly revenue of $100 billion. Management is now suggesting that growth could remain extremely high even after passing that milestone.

4. AWS ordered 2 million additional Nvidia GPUs

Amazon Web Services plans to deploy 2 million additional Nvidia GPUs across 2027 and 2028. The deployment will include Blackwell Ultra, Rubin and Rubin Ultra products.

Amazon will also use Nvidia’s physical AI platform for its warehouse robots and bring Nvidia Vera CPU-based infrastructure to AWS. The companies also plan to build US government AI infrastructure containing 100,000 GPUs.

The official AWS-Nvidia announcement provides something investors value more than general statements about strong demand: a large, identifiable customer commitment covering several product generations.

It also shows that Amazon’s internal Trainium chips do not automatically eliminate its demand for Nvidia. AWS appears to be building a mixed infrastructure model in which custom chips and Nvidia systems work alongside each other.

5. Supply covers only around 70% of current demand

Jensen Huang said Nvidia currently has enough supply to meet approximately 70% of demand, according to the Associated Press.

That statement needs a little math.

If available supply represents 70% of demand, total demand is approximately 43% higher than current supply, not simply 30% higher.

Think of a restaurant that can serve 70 customers but has 100 people waiting. Expanding the kitchen could create additional sales without needing to attract more customers.

The important risk is that Nvidia must secure enough memory, manufacturing, power and data-centre capacity before that waiting demand moves to competitors or changes its plans.

What Went Right for Nvidia in Q2 FY2027?

Blackwell Ultra delivered another large growth step

Data Center revenue increased by $13.78 billion in only one quarter, from $75.25 billion to $89.02 billion.

This accounted for approximately 94% of Nvidia’s total sequential revenue growth.

Nvidia attributed the Data Center growth primarily to the Blackwell Ultra infrastructure ramp. Data Center now contributes approximately 92.5% of total revenue, according to Nvidia’s Q2 Form 10-Q.

This concentration is both Nvidia’s greatest strength and its biggest dependency. The company is overwhelmingly an AI infrastructure business now.

Growth expanded beyond the largest cloud companies

Nvidia divides Data Center customers into two groups:

• Hyperscale: Large cloud and consumer internet platforms
• ACIE: AI clouds, industrial customers, enterprises and sovereign AI projects

ACIE generated approximately 59% of Nvidia’s sequential Data Center growth, even though it remains the smaller category.

That is one of the quarter’s most encouraging numbers. It suggests the AI infrastructure build-out is broadening beyond a handful of US technology giants.

However, ACIE does not provide perfect diversification. Nvidia says some hyperscalers also purchase capacity through AI clouds. Investors must still ask who ultimately uses and pays for the compute.

Gross margin remained at 75%

Nvidia produced approximately $72.14 billion of gross profit during the quarter while maintaining a 75% gross margin.

Adjusted operating income increased 124% year on year to $63.96 billion. Revenue more than doubled, while operating expenses increased by 54%.

That operating leverage shows the value of Nvidia’s platform model. Once the expensive research and product-development work is completed, each additional dollar of hardware and software revenue can produce considerable profit.

Vera Rubin remained on schedule

Nvidia said Vera Rubin has entered full production, with racks running at CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure and Nebius.

Production shipments began in fiscal Q3, while Nvidia built inventory ahead of the launch.

This reduces one of the biggest risks highlighted in the preview: customers do not appear to be delaying Blackwell purchases while waiting for Rubin. Blackwell Ultra is still growing rapidly even as the next platform begins production.

Nvidia achieved the growth without meaningful China sales

Data Center Hopper shipments to China contributed less than 1% of Data Center revenue in Q2. Nvidia also excluded China Data Center compute sales from its Q3 outlook.

This makes the reported growth cleaner. A recovery in China could provide upside, but the present guidance does not require it.

Nvidia nevertheless recorded a $400 million charge during the first half related to H200 excess inventory and purchase obligations as demand weakened amid regulatory restrictions. China remains an option on the outlook, not a dependable part of the base case.

What Did Jensen Huang and Nvidia Management Say?

Jensen Huang’s “compute is revenue” phrase is important, but investors should treat it as a claim that must be proven through customers’ financial results.

AI spending becomes sustainable when customers generate enough income, productivity savings or strategic value to justify the infrastructure cost. Nvidia’s sales prove that customers are ordering the equipment. They do not yet prove that every customer will earn a satisfactory return from it.

Nvidia Q3 FY2027 Guidance: What It Means

The revenue guidance was excellent, but the margin guidance explains why NVDA stock initially fell.

At $108 billion of revenue and a 74% gross margin, Nvidia would produce approximately $79.92 billion of quarterly gross profit.

That would still be around $7.78 billion more than Q2, despite the one-percentage-point margin decline.

This is the key reason investors accepted the lower margin guidance. Profit per dollar of revenue is declining slightly, but the number of revenue dollars is growing fast enough to produce substantially more total profit.

Why Nvidia’s Gross Margin Is Expected to Fall

Nvidia expects gross margin to decline from 75% in Q2 to approximately 74% in Q3. Management expects margins to bottom between 71% and 72% in Q4, before recovering towards 72% to 73% in fiscal 2028.

The main reason is the shortage and rising price of high-performance memory. Nvidia’s AI systems contain expensive HBM and other memory components, so higher memory prices increase the cost of producing each system.

Nvidia plans to increase its own prices beginning in fiscal 2028, according to MarketWatch.

At Q3’s $108 billion revenue midpoint, every one-percentage-point reduction in gross margin represents approximately $1.08 billion of quarterly gross profit.

If Nvidia’s margin falls from 75% to the midpoint of its 71% to 72% Q4 expectation, the isolated impact would equal around $3.78 billion per quarter, assuming revenue remained unchanged. Revenue is unlikely to remain unchanged, but the calculation shows why margin pressure deserves attention.

The main question is whether customers accept Nvidia’s planned price increases. Strong demand suggests Nvidia has pricing power, but Amazon, Google, Microsoft, Meta and other customers are also investing in custom chips to reduce their dependence on Nvidia.

The Sell-Earn-Collect-Commit Framework

Our Nvidia earnings preview proposed a three-part framework: how much Nvidia sells, how much it earns and how cleanly it collects the cash.

The Q2 results suggest investors now need to add a fourth test: how much Nvidia must commit to support future demand.

Sell: Demand was stronger than expected

Nvidia beat revenue consensus by $4.05 billion, Data Center revenue doubled and management projected another year of exceptional growth.

This was the cleanest part of the report.

Earn: Profit dollars are growing, but margins are falling

Nvidia preserved its 75% gross margin during Q2. The expected decline towards 71% to 72% is meaningful, but absolute gross profit could continue increasing if revenue growth remains strong.

Investors should separate lower margins caused by temporarily expensive memory from lower margins caused by weak pricing power. The first is manageable. The second would be a more serious change to Nvidia’s competitive position.

Collect: Revenue growth moved ahead of cash collection

This is the main number most earnings headlines missed.

Nvidia said receivables increased because it provided extended payment terms on large, multi-quarter agreements with investment-grade customers. Its filing says some payment periods can extend from 90 days to one year.

This does not mean Nvidia’s sales are unreliable. The company also received $15.6 billion in customer advances during the first half.

However, it means Nvidia is waiting longer to collect some of the revenue it has already recorded. Five direct customers represented 70% of total receivables at the end of the quarter.

One quarter does not establish a negative trend. But if receivables continue growing much faster than revenue, investors should reduce the weight they place on the headline earnings number.

Commit: Nvidia is making a much larger bet on future demand

Nvidia’s supply and capacity commitments increased from $119 billion to $279 billion in one quarter, primarily due to memory and manufacturing requirements.

The company separately disclosed:

These figures are not the same as current debt or immediate cash payments. They are spread across several years, some arrangements may be adjusted and guarantees only become payable under specific conditions.

Still, they show how Nvidia’s business model is changing. The company is no longer only selling AI hardware. It is securing memory, manufacturing, land, power, financing and customers’ data-centre capacity.

Think of Nvidia as a popular home builder. It has so many buyers that it is now reserving cement years in advance, helping arrange mortgages and guaranteeing parts of future housing developments.

That can allow Nvidia to build and sell more homes. It also leaves the company carrying more risk if buyers, lenders or construction partners fail to perform.

Did Nvidia’s Investment Gains Inflate Profit?

Nvidia reported GAAP net income of $59.69 billion, but this included $7.77 billion of net gains from equity investments.

That is why GAAP EPS of $2.46 was higher than adjusted EPS of $2.22. Nvidia’s adjusted result excludes these investment gains.

Investors should therefore use adjusted operating profit and adjusted EPS when analysing the performance of the core chip and computing business. The investment portfolio creates real value, but quarterly market gains should not be treated as recurring operating earnings.

What Nvidia’s Q2 Earnings Mean for NVDA Stock Investors

The results strengthen the argument that AI infrastructure demand has not yet peaked. Nvidia beat every major revenue expectation, expanded beyond hyperscalers, secured a massive AWS deployment and guided towards another exceptional growth year.

However, Nvidia’s investment thesis has become more complicated.

Two years ago, investors mainly needed to track GPU demand, competition and gross margins. They now also need to track receivables, customer payment terms, supply commitments, equity investments, data-centre leases and financial guarantees.

Our view is that this was an excellent operating quarter but only a mixed cash-quality quarter.

The market is rewarding Nvidia because the growth outlook became substantially stronger. It should not interpret the rally as proof that margin pressure, cash conversion and financing risks have disappeared.

The most encouraging signal is that demand appears limited by supply rather than customer interest. The most important warning is that Nvidia is using more of its balance sheet and financial strength to unlock that demand.

Nvidia Q2 FY2027 Earnings Analysis: Final Takeaway

Nvidia finally broke its four-quarter post-earnings losing streak because it delivered more than a routine beat.

Revenue exceeded the official guidance midpoint by $5.22 billion. Q3 guidance came in almost $4 billion above consensus. AWS committed to 2 million additional GPUs, and management projected roughly 70% revenue growth for fiscal 2028.

That was enough for investors to look past a weaker margin outlook.

The long-term question is no longer whether Nvidia can sell enough AI chips. Demand appears plentiful. The question is how much manufacturing, financing and balance-sheet risk Nvidia must accept to convert that demand into cash.

For future quarters, the best way to judge Nvidia is through all four tests: Sell, Earn, Collect and Commit.