How to Use AI to Read NVIDIA Earnings: A 5-Step Research Workflow

Intro

NVIDIA earnings are rarely hard to find.

The harder part is figuring out what actually matters.

Revenue, Data Center growth, gross margin, guidance, management commentary, after-hours moves, analyst expectations — within a few hours, there is more information than most people can reasonably process.

NVIDIA’s latest FY2027 Q2 results are a good example. Revenue reached $96.22 billion, while Data Center revenue came in at $89.0 billion, up 117% year over year. Adjusted EPS was $2.22, ahead of roughly $2.10 expected.

At first glance, the takeaway seems obvious: another strong quarter.

But earnings research gets more useful once you stop asking whether the numbers were “good” and start asking:

What changed? What drove the change? And what could challenge it next?

That is where AI can save time. It can compress the first pass, surface the numbers worth investigating, and help you keep digging without jumping between a dozen tabs.

Here is a practical way to do it.

---

1. Start With the Changes, Not the Entire Report

Reading an earnings release from top to bottom is rarely the fastest way to understand it.

Start with a narrower question:

What were the most important changes in NVIDIA’s latest quarter compared with last quarter, last year, and market expectations?

For NVIDIA, a few numbers stand out immediately:

• Revenue: $96.22B
• Data Center revenue: $89.0B
• Data Center sequential growth: roughly 18%
• Data Center year-over-year growth: roughly 117%
• Adjusted EPS: $2.22
• Market EPS expectation: roughly $2.10

Data Center now accounts for well over 90% of NVIDIA’s revenue.

More importantly, when you compare the quarter with the previous one, Data Center appears to have contributed the overwhelming majority of the incremental revenue.

So the useful takeaway is no longer:

NVIDIA grew again.

It becomes:

Why is Data Center still growing this quickly?

That is the question worth following.

---

2. Follow the Growth Back to the Customer

A 117% growth rate sounds impressive, but it does not tell you who is actually buying.

Looking deeper into NVIDIA’s earnings materials and management commentary gives a more interesting picture.

Hyperscale customers contributed roughly $49 billion in Data Center revenue, growing about 13% sequentially.

AI cloud, industrial, enterprise and sovereign AI customers contributed roughly $40 billion, growing about 25% sequentially and much faster year over year.

That shift matters.

The NVIDIA story is becoming broader than a handful of hyperscalers such as Microsoft, Amazon, Google and Meta. Demand is spreading into:

• AI cloud providers
• Model companies
• Enterprise AI
• Sovereign AI infrastructure
• Industrial computing

A broader customer base would make the growth story less dependent on a few giant buyers continuing to raise capital expenditure at the same pace.

There is another layer inside the Data Center.

Compute is still the much larger business, at roughly $71.6 billion, but Networking is growing faster. Networking revenue reached around $17.5 billion, with year-over-year growth of roughly 141%.

Products such as Spectrum-X Ethernet and NVLink are helping NVIDIA sell more of the infrastructure around the GPU itself.

The story is gradually shifting from:

selling accelerators

to:

selling more of the complete AI factory.

This was one of the more useful follow-ups when we ran the earnings through Bobby AI. Instead of stopping at the headline Data Center number, we kept asking what was changing inside it — customer mix, Compute versus Networking, and where the next layer of growth might come from.

That is often where an AI trading app becomes more useful than a simple earnings summary.

---

3. A Great Quarter Still Deserves a Bear Case

When an earnings report looks this strong, it is easy to keep collecting positive evidence.

A better research habit is to deliberately ask for the other side.

Try:

What are three risks that could be easy to overlook in this quarter?

Three areas stand out.

Supply and infrastructure commitments are rising quickly

NVIDIA’s supply and capacity commitments increased from roughly $119 billion to $279 billion.

The company has also disclosed substantial commitments related to land, power and supporting infrastructure.

That is understandable when demand is strong. NVIDIA needs capacity if customers continue building AI infrastructure at this pace.

But it also means the company is making much larger commitments against future demand.

If deployment slows, utilization disappoints, or product cycles shift faster than expected, those commitments become more important to watch.

Cash conversion was weaker this quarter

Operating performance remained strong, but free cash flow did not keep pace with earnings growth.

Quarterly operating cash flow was roughly $24.08 billion, with free cash flow around $21.4 billion after capital expenditures.

That was well below the prior quarter’s roughly $48.6 billion in free cash flow.

One quarter is not a trend. Some of the gap reflects inventory, working capital and investment required to support the next phase of AI infrastructure.

Still, if profits continue to rise much faster than cash generation over several quarters, it deserves attention.

Gross margin may be entering a different phase

NVIDIA reported a GAAP gross margin of roughly 75% for Q2.

Its Q3 guidance points to around 74% ± 0.5 percentage points.

A 74% gross margin is hardly weak.

The interesting part is the direction.

The next phase of growth may require more expensive memory, full-rack systems, networking equipment and increasingly complex infrastructure.

The story may therefore become less:

revenue up, margins up

and more:

revenue still growing quickly, but some of that growth is becoming more expensive to deliver.

This is why asking Bobby AI for both a bull case and a bear case is more useful than asking whether the report was simply “good.”

AI does not need to choose the side for you. It is often more valuable when it helps you see the strongest evidence on both sides.

---

4. Good Earnings Do Not Guarantee an Immediate Rally

NVIDIA’s post-earnings move is a useful reminder that an earnings report and the market’s reaction to it are two different things.

The stock closed around $209.66 before the report.

Immediately after the release, shares initially fell close to 2%.

Then the reaction changed.

As investors digested management commentary and guidance, the stock reversed higher. By the following session, NVIDIA closed around $227.98, roughly 8.7% above its pre-earnings close.

Trading volume also jumped well above its recent average.

Why the hesitation?

Investors were processing several things at once:

• Revenue came in above expectations.
• Data Center remained exceptionally strong.
• Q3 revenue guidance was around $108 billion, above the roughly $104.2 billion market expectation.
• The guidance still excluded China Data Center compute revenue.
• Gross margin is expected to moderate.
• AI infrastructure spending commitments continue to rise.

The market was not simply deciding whether the quarter was “good.”

It was repricing the next stage of NVIDIA’s growth.

That distinction is useful when looking at stock trading signals around earnings.

A momentum signal, volume spike or technical breakout tells you something changed. It does not explain what changed, or whether the first reaction will last.

For earnings-driven moves, context matters as much as the signal itself.

---

5. Earnings Research Should End With a Watchlist, Not a Verdict

A useful earnings report should tell you what to watch next.

For NVIDIA, three areas deserve attention over the coming quarter.

Data Center customer mix

The headline Data Center growth rate still matters, but the composition may become increasingly important.

If AI cloud, enterprise and sovereign AI customers continue to outgrow traditional hyperscalers, NVIDIA’s demand base is becoming more diversified.

Gross margin

Revenue can continue growing while the economics of delivering that revenue change.

If gross margin keeps moving down from current levels, the next question is why:

Is it simply product mix, or is the cost of building and delivering full AI systems structurally increasing?

Cash conversion

If earnings continue to grow rapidly while free cash flow lags, inventory, working capital and long-term supply commitments become increasingly important.

Those questions are arguably more useful than asking:

Will NVIDIA beat again next quarter?

---

6. Bring NVIDIA Back to Your Own Portfolio

Once the company-level research is done, the question can become more personal.

Instead of:

Is NVIDIA a good company?

Ask:

If I already own NVIDIA, AMD and several other AI stocks, how concentrated is my portfolio really?

Or:

NVIDIA and AMD are different companies, but am I effectively betting on the same AI capex cycle twice?

Or simply:

If I make no trade at all, what should I monitor before the next earnings report?

This is where TradeGPT-style tools and newer AI trading app workflows become more useful.

The research does not have to stop at one company. It can move from:

earnings → industry → competitors → portfolio

without starting over each time.

With Bobby AI, for example, the same NVIDIA conversation can continue into company comparisons or portfolio questions inside the RockFlow ecosystem.

That does not mean handing the final decision to AI.

The same principle applies to automated trading and CopyTrading. Automation can reduce repetitive execution, and CopyTrading can make an existing strategy easier to follow. Neither removes the need to understand what an earnings event has changed underneath the strategy.

---

A Prompt You Can Reuse for NVIDIA Earnings

If you want a cleaner starting point next time, try this:

Analyze NVIDIA’s latest earnings report, but do not tell me whether I should buy or sell the stock.

Please:

1. Identify the most important financial changes.
2. Compare them with the previous quarter, the same quarter last year, and market expectations.
3. Explain which businesses and customer groups drove the growth.
4. Identify three risks that could be easy to overlook.
5. Explain the market reaction after the report.
6. Give me three metrics worth tracking over the next quarter.

Cite the source for important numbers and clearly separate facts, market views and inference.

Then keep going.

If Data Center looks interesting:

Go deeper into the customer mix.

If the first answer sounds overwhelmingly positive:

Now make the strongest bear case from the same data.

When we tested this workflow with Bobby AI, the most useful part was not the first summary. It was the follow-up questions.

The first answer usually tells you what deserves a second question.

---

FAQ

Can AI tell me whether I should buy NVIDIA?

AI is more useful for research than for making the final decision for you.

It can organize earnings data, compare periods, surface risks and explain market reactions. Whether to trade, how much to invest and how much risk to take still depend on your own circumstances.

What is the difference between a general AI assistant and an AI trading app?

A general AI assistant is useful for explaining documents and organizing information you provide.

An AI trading app may also connect market data, company research, portfolio information or trading workflows.

Capabilities vary by product, so it is still important to understand where the data comes from and how current it is.

Are stock trading signals useful after earnings?

They can be useful as an alert that something changed.

A stock trading signal does not explain the full reason behind a move. Earnings, guidance, industry performance and market expectations may all contribute at the same time.

Are TradeGPT, automated trading and CopyTrading the same thing?

No.

TradeGPT generally refers to using GPT-style natural-language interaction in investing or trading workflows.

Automated trading focuses on executing predefined rules automatically.

CopyTrading focuses on following an existing trader or strategy.

They can exist within the same platform, but they solve different parts of the investing workflow.

---

Final Thoughts

• This quarter, the more interesting story may not simply be that NVIDIA is approaching $100 billion in quarterly revenue.
• It is that the customer mix inside Data Center is broadening, Networking is growing faster than Compute, infrastructure commitments are rising sharply, and both gross margin and cash conversion are beginning to show a slightly different shape.
• AI can make that research process faster:

Find the change → understand the driver → look for the risk → check the market reaction → decide what to watch next.

• If you want to run the same process yourself, start with NVIDIA’s earnings report in Bobby AI and keep following whichever part of the answer looks incomplete. The goal is not to get AI to say “buy” or “sell.” It is to get to the next useful question faster.