Nvidia – What’s next?
NVIDIA Earnings: AI Demand Surges Again — What Wall Street Is Saying and Which AI Stocks Benefit Next
NVIDIA has once again delivered the kind of earnings report that can reshape the entire technology sector.
The company reported $96.2 billion in quarterly revenue, up more than 100% year over year, while Data Center revenue reached an extraordinary $89 billion, up 117%. Adjusted EPS came in at $2.22, versus approximately $2.09 expected by analysts. For the next quarter, NVIDIA guided toward roughly $108 billion of revenue. (AP News)
But the biggest takeaway may not be the quarter itself.
NVIDIA’s management indicated that revenue could still grow by approximately 70% in fiscal 2028, suggesting that the current AI infrastructure cycle is far from reaching maturity. (Axios)
For investors, this creates a much bigger question:
What does NVIDIA’s earnings report mean for the rest of the AI ecosystem?
The answer is increasingly clear: the AI trade is broadening beyond GPUs.
📊 NVIDIA Earnings at a Glance
| Metric | Q2 FY2027 | YoY / Comparison |
|---|---|---|
| Revenue | $96.22B | +107% |
| Adjusted EPS | $2.22 | vs. $2.09 expected |
| Data Center Revenue | $89.0B | +117% |
| Q3 Revenue Guidance | ~$108B | ~89% YoY implied |
| FY2028 Growth Outlook | ~70% | Extremely strong |
| After-hours reaction | ~+4% | Strongly positive |
NVIDIA’s Data Center business now generates more revenue in a quarter than many major semiconductor companies generate in an entire year.
That scale is what makes this earnings report so important for the entire market. (Reuters)
🚀 The Most Important Number Isn’t $96 Billion
The headline $96.2 billion revenue figure is spectacular.
But investors should focus even more heavily on the forward trajectory.
NVIDIA expects approximately $108 billion of revenue in Q3, while management sees roughly 70% revenue growth in fiscal 2028. (Axios)
This matters because the primary bear thesis surrounding AI has been:
AI infrastructure spending is approaching a peak.
NVIDIA’s guidance argues the opposite.
The company is effectively telling investors:
Blackwell is still ramping.
Rubin is beginning to ramp.
Inference is becoming increasingly important.
AI workloads continue to expand.
Customers remain supply constrained.
And potentially most importantly:
NVIDIA still sees enormous infrastructure demand beyond the current product cycle.
🧠 Jensen Huang: AI Has Entered a New Phase
CEO Jensen Huang characterized the current environment as a pivotal period for AI, with NVIDIA increasingly positioning itself as a complete AI infrastructure platform rather than simply a GPU supplier.
That distinction matters.
NVIDIA is increasingly involved in:
- AI training
- AI inference
- networking
- CPUs
- complete AI systems
- software
- AI models
- data centers
- infrastructure financing
- ecosystem investments
MarketWatch’s analysis highlighted precisely this evolution, noting that NVIDIA is moving beyond training into inference, models, networking and broader infrastructure. (MarketWatch)
This significantly expands NVIDIA’s addressable market.
🔥 Wall Street Analyst Reaction
The analyst reaction is particularly interesting because the bar was already extremely high.
Before the report, roughly 60 analysts were forecasting around $91.9 billion of revenue and $2.09 EPS. (IG)
NVIDIA exceeded both.
But analysts aren’t simply reacting to the quarter.
They are increasingly focusing on Blackwell, Rubin, inference, free cash flow, AI infrastructure spending and the sustainability of hyperscaler capex.
🟢 UBS: NVIDIA Could Still Be Heading Toward $20 EPS
UBS has been one of the more bullish voices on NVIDIA.
Ahead of earnings, UBS expected approximately $94–95 billion of quarterly revenue and projected that the following quarter could exceed $110 billion.
More importantly, UBS argued that investors could gain greater confidence in a path toward approximately:
$15+ EPS in calendar 2027
and
$20 EPS in calendar 2028. (Business Insider)
That is an extremely important observation.
If NVIDIA can approach $20 EPS in 2028, then the investment debate changes from:
“Is NVIDIA expensive?”
to:
“What multiple should investors apply to $20+ of future earnings?”
That is a much more constructive framework.
🟢 Bank of America: $350 Price Target and Bigger Buybacks
Bank of America analyst Vivek Arya remains strongly bullish on NVIDIA.
His price target was $350.
But Arya also raised another interesting issue: NVIDIA’s enormous free cash flow could support substantially greater shareholder returns.
He estimates that NVIDIA could eventually generate close to $1 billion in cash per day, allowing the company to invest heavily in its ecosystem while simultaneously increasing buybacks. (MarketWatch)
BofA’s argument is essentially:
NVIDIA doesn’t have to choose between investing for growth and returning capital.
It may be able to do both.
That becomes increasingly relevant as NVIDIA’s cash generation scales.
🟢 KeyBanc: Rubin Is the Next Major Catalyst
One of the most important analyst themes going into earnings was the Vera Rubin ramp.
That thesis was validated.
NVIDIA says Rubin is entering full production and is already being deployed by major customers including:
- Microsoft
- Google Cloud
- Oracle
- CoreWeave
- Nebius
This is important because it means NVIDIA’s growth isn’t dependent entirely on Blackwell.
Instead, investors are beginning to see a succession of AI computing platforms:
Blackwell → Rubin → Future Architectures
That creates the potential for a much longer semiconductor cycle than investors experienced during traditional PC or smartphone upgrades.
🟡 Bernstein: Watch the Financing Model
Bernstein’s Stacy Rasgon has been one of the more important voices questioning NVIDIA’s broader ecosystem strategy.
The concern isn’t really whether AI demand exists.
It clearly does.
The bigger question is:
How much of the future AI infrastructure buildout depends on financing structures involving NVIDIA itself?
NVIDIA has increasingly participated in AI infrastructure financing and ecosystem investments.
That creates the possibility of what investors sometimes call circular financing:
NVIDIA invests → infrastructure gets built → customers buy NVIDIA hardware → NVIDIA revenue rises → NVIDIA has more capital to invest.
That doesn’t mean the revenue isn’t real.
But it does mean investors need to examine the economic returns generated by the infrastructure, not just the amount of hardware being purchased.
NVIDIA has sought to reassure investors on this issue, including disclosure that its maximum gross exposure to certain land, power and shell guarantees was approximately $3.5 billion. (Reuters)
For now, this remains a risk to monitor rather than evidence that the AI boom is artificial.
🟡 Morgan Stanley: What Happens With Open-Source AI?
Morgan Stanley’s analyst questioning focused on an increasingly important issue:
How does NVIDIA perform as AI models become cheaper and more widely available?
This is actually a bullish question for the overall AI ecosystem.
If open-source models become cheaper to use, that could increase the number of:
- AI applications
- inference workloads
- AI agents
- enterprise deployments
- AI searches
- autonomous systems
In other words:
Cheaper AI does not necessarily mean less compute.
It could mean much more compute consumption.
And NVIDIA is positioned throughout that infrastructure.
🟢 Analyst Consensus: The Bull Case Remains Dominant
The pre-earnings setup was already extraordinarily bullish.
One pre-earnings estimate showed roughly 79 of 82 analysts carrying Buy ratings, illustrating just how overwhelmingly positive Wall Street sentiment was. (Business Insider)
That creates an interesting paradox.
NVIDIA is overwhelmingly loved by analysts.
But expectations are also extraordinarily high.
That’s why NVIDIA has sometimes sold off even after beating estimates.
In fact, the stock had declined after four of its previous five earnings reports despite beating EPS expectations each time. (Yahoo Finanzen)
This time, however, the combination of:
Huge beat + $108B guidance + Rubin ramp + long-term growth visibility
was strong enough to push shares approximately 4% higher after the report. (Reuters)
🌐 The Biggest Investment Implication: AI Is Becoming a Multi-Layer Trade
The NVIDIA earnings report provides validation for far more than GPUs.
The AI infrastructure chain increasingly looks like this:
| Layer | Key Companies | NVIDIA Read-Through |
|---|---|---|
| 🧠 AI Accelerators | NVDA, AMD | ⭐⭐⭐⭐⭐ |
| ⚙️ Custom AI Silicon | AVGO, MRVL | ⭐⭐⭐⭐⭐ |
| 🧮 HBM / Memory | MU | ⭐⭐⭐⭐⭐ |
| 🏭 Foundry | TSM | ⭐⭐⭐⭐⭐ |
| 🌐 Networking | AVGO, MRVL, ANET | ⭐⭐⭐⭐⭐ |
| 💾 Storage | SNDK | ⭐⭐⭐⭐½ |
| 🖥️ AI Servers | DELL, SMCI | ⭐⭐⭐⭐ |
| ☁️ AI Cloud | CRWV, NBIS, ORCL | ⭐⭐⭐⭐ |
| 🤖 AI Software | PLTR + software | ⭐⭐⭐ |
This is the most important evolution in the AI investment thesis.
🟢 AMD: NVIDIA Validates the Accelerator Market
AMD is arguably one of the clearest beneficiaries.
NVIDIA’s $89 billion Data Center quarter demonstrates the extraordinary size of the accelerator market.
AMD doesn’t need to replace NVIDIA.
It needs to capture incremental market share.
And if the overall accelerator market continues expanding at extraordinary rates, AMD can grow rapidly even while NVIDIA maintains its dominant position.
StockInsight™ View: 🟢 Bullish
Primary catalyst: AI accelerator market expansion.
Primary risk: NVIDIA’s software ecosystem and architectural lead remain enormous.
🟢 Broadcom: One of the Biggest Second-Order Winners
Broadcom may ultimately prove to be one of the most important beneficiaries of the AI boom outside NVIDIA.
Why?
Because hyperscalers increasingly want:
custom AI accelerators.
Google has TPUs.
Amazon has Trainium and Inferentia.
Microsoft is developing its own AI silicon.
Meta is developing custom accelerators.
The larger AI infrastructure becomes, the more economically attractive custom silicon becomes.
That creates an enormous opportunity for Broadcom.
StockInsight™ View: 🟢🟢 Very Bullish
AI thesis: Custom silicon + networking.
Key catalyst: Hyperscaler AI ASIC deployments.
🟢 Marvell: A Major AI Infrastructure Read-Through
Marvell is another company investors should watch closely.
The company’s exposure includes:
- custom silicon
- networking
- connectivity
- optical infrastructure
- data-center infrastructure
That makes it a classic second-order NVIDIA beneficiary.
The stock moved higher following NVIDIA’s earnings, reflecting the market’s expectation that AI infrastructure spending remains robust. (Reuters)
StockInsight™ View: 🟢🟢 Very Bullish
Key question: Can Marvell convert AI infrastructure demand into sustained revenue and margin growth?
🟢🟢 Micron: Memory Is Becoming a Bottleneck
This may be one of the most important takeaways from the report.
NVIDIA highlighted extreme memory-cost inflation.
For NVIDIA, that’s a margin headwind.
For memory suppliers, it is evidence of:
scarcity + pricing power + strong AI demand.
NVIDIA expects gross margins to come under pressure as memory costs rise, with estimates around the 72–73% range in fiscal 2028 compared with approximately 75% currently. (Business Insider)
That creates an interesting divergence:
NVDA: higher memory costs = negative
MU: higher memory prices = potentially positive
This is exactly the kind of second-order AI trade that investors should watch.
🟢 TSMC: The Foundry Bottleneck
If NVIDIA, AMD, Broadcom and other AI-chip designers continue increasing production, someone has to manufacture those chips.
That makes TSMC one of the fundamental beneficiaries of the AI infrastructure cycle.
The important question isn’t whether AI demand is strong.
NVIDIA has effectively answered that.
The question becomes:
Can the semiconductor supply chain manufacture enough advanced AI silicon to meet demand?
That creates a favorable long-term setup for advanced-node foundry capacity.
🟢 Arista Networks: AI Networking Becomes Critical
AI clusters are becoming enormous.
Thousands — and eventually hundreds of thousands — of accelerators need to communicate with one another.
That makes networking just as important as compute.
As AI clusters scale, high-speed Ethernet and networking infrastructure become increasingly critical.
Therefore, NVIDIA’s growth outlook provides a positive read-through for Arista Networks.
StockInsight™ View: 🟢 Bullish
The AI networking opportunity may ultimately become one of the largest second-order beneficiaries of the AI buildout.
🟢🟢 CoreWeave and Nebius: The AI Cloud Trade
The market reaction here was particularly interesting.
Following NVIDIA’s earnings, AI-cloud companies were among the strongest beneficiaries.
That’s logical.
If NVIDIA sells enormous quantities of GPUs, someone has to deploy them.
That is where companies such as CoreWeave and Nebius become relevant.
The recent earnings cycle had already demonstrated strong demand for AI compute, with CoreWeave reporting rapidly growing revenue and a substantial backlog. (Business Insider)
NVIDIA’s earnings provide another layer of validation.
But there is a major caveat.
AI cloud companies have:
- enormous capital requirements
- high depreciation
- financing requirements
- customer concentration
- infrastructure execution risk
Therefore:
AI demand ≠ automatically attractive AI-cloud economics.
This distinction is critical.
⚠️ The Biggest Risk to the AI Trade
The biggest risk isn’t that NVIDIA’s technology suddenly becomes obsolete.
It is that AI infrastructure spending eventually produces inadequate returns on capital.
The industry is now spending hundreds of billions of dollars on:
- GPUs
- data centers
- electricity
- networking
- cooling
- memory
- storage
- fiber
- AI models
The question investors eventually need answered is:
How much economic value will all of this infrastructure generate?
Microsoft and Meta alone are expected to contribute to more than $730 billion of AI infrastructure spending in 2026, according to Reuters’ reporting. (Reuters)
That is an extraordinary amount of capital.
For now, NVIDIA’s earnings suggest demand is real.
But eventually the market will demand proof of AI ROI.
⚠️ NVIDIA’s Margin Story Is Changing
This is another issue investors shouldn’t overlook.
NVIDIA’s revenue growth remains extraordinary.
But margins are facing pressure.
The primary issues include:
- higher memory prices
- product mix
- manufacturing costs
- increasingly complex AI systems
- infrastructure investments
NVIDIA’s expected gross-margin normalization toward approximately 72–73% by FY2028 doesn’t invalidate the growth story. (Business Insider)
But it does mean investors should stop thinking of NVIDIA as a company that can simultaneously deliver:
100% growth + 75% margins + unlimited operating leverage
forever.
Eventually, growth will normalize.
The question is how high the earnings base becomes before that happens.
📈 StockInsight™ AI Beneficiary Ranking
| Rank | Stock | AI Earnings Read-Through | Outlook |
|---|---|---|---|
| 🥇 | NVDA | Direct AI compute leader | 🟢🟢🟢🟢🟢 |
| 🥈 | AVGO | Custom AI + networking | 🟢🟢🟢🟢🟢 |
| 🥉 | AMD | Accelerator market expansion | 🟢🟢🟢🟢🟢 |
| 4 | MU | HBM / memory pricing | 🟢🟢🟢🟢🟢 |
| 5 | MRVL | Custom silicon + connectivity | 🟢🟢🟢🟢🟢 |
| 6 | TSM | Advanced-node manufacturing | 🟢🟢🟢🟢🟢 |
| 7 | ANET | AI networking | 🟢🟢🟢🟢½ |
| 8 | SNDK | AI storage | 🟢🟢🟢🟢½ |
| 9 | CRWV | AI compute deployment | 🟢🟢🟢🟢 |
| 10 | NBIS | AI cloud | 🟢🟢🟢🟢 |
| 11 | ORCL | AI cloud infrastructure | 🟢🟢🟢🟢 |
| 12 | DELL | AI server demand | 🟢🟢🟢½ |
| 13 | SMCI | AI servers | 🟢🟢🟢½ |
| 14 | PLTR | AI application layer | 🟢🟢🟢 |
🔮 Bull Case for the AI Trade
The bullish scenario is becoming increasingly powerful.
🟢 1. Hyperscaler CAPEX continues rising
Microsoft, Meta, Google and Amazon continue spending aggressively on AI infrastructure.
🟢 2. Rubin extends NVIDIA’s product cycle
Blackwell isn’t the end of the current AI hardware cycle.
Rubin creates another major upgrade cycle.
🟢 3. Inference explodes
Training has dominated the first phase of AI.
Inference could become the much larger long-term workload as AI agents and enterprise applications proliferate.
🟢 4. AI expands beyond GPUs
Networking, memory, storage, custom silicon and power infrastructure all benefit.
🟢 5. AI becomes cheaper
Ironically, cheaper AI could increase demand because more businesses can afford to deploy it.
🔴 Bear Case
There are still substantial risks.
🔴 AI ROI disappoints
Companies spend hundreds of billions but fail to generate sufficient incremental revenue.
🔴 Hyperscaler CAPEX slows
If Microsoft, Google, Amazon or Meta reduce spending, the entire ecosystem would feel it.
🔴 Custom silicon takes share
NVIDIA’s largest customers increasingly develop their own accelerators.
🔴 Margins compress
Memory and infrastructure costs remain elevated.
🔴 Financing concerns grow
The market becomes increasingly concerned about AI infrastructure being financed through interconnected ecosystem relationships.
🔴 Valuation becomes the problem
Even exceptional earnings growth can produce disappointing stock returns if expectations become too high.
🎯 What Investors Should Watch Next
The next phase of the AI trade will be about confirmation.
Watch these companies closely:
AMD
Does AI accelerator revenue continue accelerating?
Broadcom
Are custom AI ASIC deployments expanding?
Marvell
Are custom silicon and AI networking accelerating?
Micron
Does HBM pricing remain strong?
TSMC
Does advanced-node capacity remain constrained?
Arista
Does AI networking continue outperforming traditional networking?
CoreWeave / Nebius
Can AI compute demand translate into attractive economics?
Oracle
Can AI infrastructure translate into sustainable cloud growth?
These companies will tell investors whether NVIDIA’s extraordinary numbers represent a company-specific phenomenon or a much broader AI capital-expenditure cycle.
🧩 The Bigger Picture
NVIDIA’s earnings have effectively answered one of the biggest questions facing markets in 2026:
Is AI infrastructure demand collapsing?
For now, the answer is clearly:
No.
Quite the opposite.
NVIDIA’s Data Center revenue is growing at triple-digit rates, Q3 guidance points toward another enormous sequential increase, and management still sees approximately 70% revenue growth in FY2028. (Reuters)
The more interesting question has therefore changed.
It is no longer:
“Is the AI boom real?”
It is:
“Who captures the next $1 trillion of AI infrastructure spending?”
That is where the investment opportunity becomes much broader.
💡 StockInsight™ Final Take
NVIDIA’s earnings are a major fundamental validation of the AI infrastructure cycle.
The company has once again demonstrated that AI demand remains extraordinary — and the forward outlook suggests that this isn’t simply a one-year spending surge.
The biggest opportunity may now be in the second and third derivatives of AI spending.
NVDA remains the dominant AI compute franchise.
But the earnings report strengthens the investment case for:
AVGO → custom silicon + networking
AMD → alternative AI accelerators
MRVL → AI connectivity + custom silicon
MU → HBM and memory
TSM → advanced manufacturing
ANET → AI networking
SNDK → AI storage
CRWV / NBIS → AI compute deployment
The key risk is no longer whether NVIDIA can sell GPUs.
It clearly can.
The key risk is whether the hundreds of billions of dollars being invested in AI infrastructure will eventually generate sufficient economic returns.
For now, NVIDIA’s numbers strongly support the bullish side of that debate.
And that makes this earnings report much more than another NVIDIA beat.
It is another major confirmation that the AI infrastructure supercycle is still expanding. 🚀
📌 StockInsight™ Key Takeaways
🟢 NVIDIA: Fundamental momentum remains exceptional
🟢 AMD: Huge accelerator TAM remains available
🟢 AVGO: Custom AI silicon increasingly attractive
🟢 MRVL: Strong second-order AI infrastructure beneficiary
🟢 MU: Memory scarcity creates pricing power
🟢 TSM: Advanced-node demand remains structurally strong
🟢 ANET: AI networking becomes increasingly critical
🟢 SNDK: AI storage demand expanding
🟢 CRWV/NBIS: AI compute demand validated, but financing risk higher
🟡 AI Software: Long-term opportunity strong, but monetization remains the next test
StockInsight™ Verdict: 🟢 BULLISH ON AI INFRASTRUCTURE — WITH SELECTIVITY
