There’s a strange, almost mythic quality to Nvidia’s place in the stock market right now. It’s the kind of phenomenon that Wall Street professors will assign as a case study for the next 40 years: the company that became the oxygen supply of the AI revolution — a company that seemed too loved, too expensive, too parabolic — suddenly becomes… cheap?
Cheap?
Yes.
Cheap.
Not “bargain-bin, rummage-sale cheap,” but “shockingly reasonable given its dominance and growth runway” cheap. And not compared to the dusty corners of the S&P 500, but compared to the Magnificent Seven — that elite tech aristocracy that investors watch as if each one were a celestial deity.
The Mag 7 used to be Apple, Microsoft, Nvidia, Alphabet, Amazon, Meta, and Tesla — a lineup of heavyweights so dominant that economists began worrying whether antitrust regulators needed emotional support animals. But even inside that royal court, Nvidia is now something surprising:
The second cheapest stock of the group.
Not by share price — which means nothing.
Not by hype — which Nvidia has in industrial quantities.
But by valuation metrics like forward P/E relative to earnings growth, free cash flow expansion, profitability, and long-term structural demand.
In other words, the stock that bears spent three years insisting “must crash” has quietly marched into the value aisle.
Let’s break that down — what it means, why it matters, and where Nvidia goes from here.
I. The Mag 7 Reset: When “Cheap” Starts to Mean Something Different
After the market reset of 2024–2025, the Magnificent Seven separated into two camps:
Camp A: Still-Great, But Fully Priced
-
Microsoft
-
Apple
-
Alphabet
-
Amazon
These companies are still strong, still massive, still churning out cash — but their valuations reflect slow(er) growth, regulatory ceilings, and the gravitational pull of sheer size.
Camp B: The Growth Rockets
-
Nvidia
-
Meta
-
Tesla (well… in theory)
But something happened in 2025: Nvidia’s earnings detonated into a new plane of reality. Meta’s spending discipline turned it into a cash geyser. Tesla… continued to do Tesla things.
When earnings grow far faster than price, valuations compress. And compress Nvidia’s did — dramatically.
Now, relative to the Mag 7:
-
Apple trades at a premium to its growth
-
Microsoft trades at a premium to its cloud dominance
-
Amazon trades at a premium to its cash-flow optionality
-
Alphabet trades at a premium to its ad monopoly
-
Meta trades cheaply, but for reasons (regulatory, political, reputational)
-
Tesla trades in its own universe, with its own physics
-
Nvidia trades at a valuation below nearly all of them except Meta — while growing faster than the entire group combined
That’s not a typo.
Nvidia is cheaper than Apple, Microsoft, Amazon, Alphabet, and Tesla on growth-adjusted valuation metrics.
Let’s get specific.
II. The Actual Numbers: Nvidia’s Valuation Doesn’t Match Its Growth (In a Good Way)
The mainstream narrative is still anchored in late-2023 thinking:
“Nvidia can’t possibly grow fast enough to justify its price.”
But then Nvidia did the most Nvidia thing imaginable:
It grew even faster.
And not by a little.
By triple-digit percentages.
Repeatedly.
Across consecutive quarters.
Across multiple business lines.
The P/E Shock
At the beginning of the AI boom, Nvidia’s forward P/E hovered above 60–70.
Today?
It’s often in the mid-20s to low-30s, fluctuating with quarter-to-quarter beats.
Which puts Nvidia’s valuation at:
-
Lower than Microsoft
-
Lower than Amazon
-
Lower than Tesla
-
Lower than Apple
-
Lower than Alphabet
-
Often only slightly above Meta
And unlike several members of this elite group, Nvidia’s earnings aren’t cyclical, flattening, or slow-growing.
They’re still explosive.
Price-to-Earnings Growth (PEG): The “Are You Actually Cheap?” Test
PEG ratio compares valuation to growth.
A PEG of:
-
2.0 = normal for megacap tech
-
1.5 = attractive
-
1.0 = undervalued
-
Below 1.0 = “Are we sure this spreadsheet isn’t wrong?”
Nvidia often hovers close to 1.0, sometimes below it, depending on the period.
This is absurd for a company generating:
-
monopolistic margins
-
record-breaking free cash flow
-
demand so high customers place orders years in advance
-
pricing power so intense it makes luxury brands jealous
The only other Mag 7 member with a PEG below 1.0 is Meta, but Meta’s margins are not Nvidia’s margins, and Meta controls eyeballs — Nvidia controls the AI industrial supply chain.
This matters.
This changes the narrative.
This is why the phrase “2nd cheapest” isn’t clickbait — it’s math.
III. Why Nvidia Deserves Its Dominance — And Why It Isn’t Going Away
People keep waiting for “the AI GPU bubble” to pop.
But what if the bubble isn’t a bubble?
What if it's a multi-decade infrastructure cycle?
Let’s look at why Nvidia is still the nucleus of global compute.
1. The AI Compute Demand Curve Isn’t Slowing — It’s Steepening
Training LLMs doesn’t get cheaper.
Inference demand doesn’t flatten.
Models grow, and they compound.
Inside every tech company boardroom:
“We need more compute.”
Inside every startup pitch deck:
“We need more compute.”
Inside every government R&D budget:
“We need sovereign compute.”
Inside every Fortune 500 CIO report:
“We need GPU clusters yesterday.”
AI isn’t a moment.
It’s a megatrend.
And Nvidia is the tollbooth.
2. Nvidia Still Owns 80–90% of the AI GPU Market
Competitors are trying:
-
AMD: decent, but still chasing CUDA
-
Intel: trying, again
-
hyperscalers: designing custom silicon, but still buying Nvidia
-
governments: hoping open-source models reduce reliance
None of them come close — because the hardware alone isn’t the product.
The product is:
CUDA + cuDNN + TensorRT + the entire Nvidia software labyrinth that developers are locked into.
Hardware is replaceable.
Ecosystems are not.
3. Nvidia’s Transition from Hardware to Platform Is Complete
Nvidia is no longer:
-
a GPU company
-
a chip company
-
a gaming company
-
a data center company
It is:
a vertically integrated compute ecosystem.
Think:
-
chips
-
systems
-
software
-
networking
-
cloud services
-
simulation environments
-
robotics stacks
-
developer platforms
-
inference engines
-
enterprise AI frameworks
Every piece strengthens the moat.
4. Nvidia’s “Competitors” Are Actually Its Customers
Microsoft?
Buys billions in Nvidia GPUs.
Amazon?
Buys billions in Nvidia GPUs even though it has Trainium.
Google?
Buys GPUs even with TPU v6.
Meta?
Probably the single most Nvidia-dependent company on earth.
Tesla?
Uses Nvidia in areas they don’t advertise loudly.
Even if the hyperscalers build custom chips, they still buy Nvidia hardware.
That’s not competition.
That’s recurring revenue.
IV. The World Is Just Beginning the AI Buildout — And Nvidia Is the Pickaxe Vendor
Think the AI boom is mature? No.
We’re barely in Year Three of what will likely be a 20-year cycle.
Here’s the roadmap:
Phase 1 (2023–2025): LLM Training Arms Race
-
ChatGPT
-
Gemini
-
Claude
-
Llama
-
Open-source explosion
-
Model scaling wars
This phase required massive GPU clusters.
Phase 2 (2025–2030): Inference Explosion
Inference is the real money printer.
Every app.
Every website.
Every enterprise workflow.
Every autonomous system.
Every industry.
Inference demand will dwarf training demand.
More GPUs. More network fabric. More everything.
Phase 3 (2030+): AI Everywhere
AI does not peak.
It becomes infrastructure.
Like electricity.
Like the internet.
Like cloud computing.
And Nvidia is not merely participating — it’s architecting.
V. The Margins Wall Street Still Underestimates
Nvidia’s margin profile is…
how do we put this politely…
obscene.
Gross margins: often >75%
Operating margins: often >60%
Free cash flow margins: industry-defying
You don’t get margins like that unless you have:
-
monopoly-level demand
-
zero pricing pressure
-
zero real substitutes
-
a captive developer base
-
a luxury-brand moat inside the enterprise market
Nvidia isn’t just selling chips.
It’s selling time.
Time to train models.
Time to deploy products.
Time to innovate faster than competitors.
In business, time is priceless — and Nvidia monetizes it.
VI. But Isn’t the Competition Going to Kill Nvidia?
Ah, the eternal bear argument.
Let’s evaluate the usual suspects.
1. AMD
Great engineers.
Solid hardware.
But missing one ingredient:
CUDA.
Without it, AMD’s gains will remain incremental — not existential threats.
2. Google, Amazon, Microsoft (Custom ASICs)
Yes, hyperscalers are designing chips.
Yes, they want to reduce Nvidia dependency.
But consider:
-
Custom ASICs are highly specialized
-
They don’t replace general-purpose GPUs
-
They depend on specific workloads
-
They require ecosystems the hyperscalers don’t supply to the market
-
They still buy Nvidia hardware at enormous scale
Building an internal chip doesn’t eliminate external need.
It reduces vendor concentration — not vendor relevance.
3. China’s Domestic Chips
Useful for sovereignty.
Not remotely on par with Nvidia.
Not exportable globally.
Not competitive in performance per watt.
Years behind on software ecosystems.
4. Regulatory Headwinds
The U.S. exports restrictions reduce Nvidia’s China revenue — yes.
But here’s the twist:
Those same restrictions force China to increase domestic investment in training infrastructure, which increases global demand for AI model innovation elsewhere, which increases global orders for Nvidia.
Restricting Nvidia in one market expands its necessity in another.
VII. Why Nvidia Became the 2nd Cheapest: Earnings Caught Up with the Hype
Simply put:
-
The stock didn’t fall
-
The earnings surged
-
The valuation compressed
Wall Street was slow to accept that:
Nvidia is not a cyclical semiconductor.
Nvidia is the backbone of the post-cloud era.
And when a company’s earnings grow 100–200% multiple years in a row, even a skyrocketing stock price can become fundamentally cheap.
VIII. Comparing Nvidia to the Other Mag 7 (This Is Where It Gets Fun)
Let’s look at a growth/valuation snapshot:
| Company | Fwd P/E | Growth Outlook | PEG | Profit Margins | AI Exposure |
|---|---|---|---|---|---|
| Apple | ~28–32 | Low | High | Medium | Weak |
| Microsoft | ~33–38 | Medium | Elevated | High | Strong |
| Amazon | ~45–60 | Medium | High | Low | Moderate |
| Alphabet | ~25–28 | Medium | Moderate | High | High |
| Meta | ~20–22 | High | Low | High | Growing |
| Tesla | ~60–70+ | Uncertain | Absurd | Low–Medium | Moderate |
| Nvidia | 25–35 | Very High | Low | Ultra High | Maximal |
Now reread that table and try telling me Nvidia shouldn’t trade at the highest multiple.
Instead, it trades near the second lowest.
It is mathematically incorrect.
Fundamentally illogical.
And historically rare.
This is not normal behavior for a megacap growth engine.
It is a generational mispricing.
IX. The Risk Factors (Because This Is Still a Fair Fight)
Let’s be honest about risks:
1. AI demand slows
Possible.
But unlikely.
Even in recessionary environments, enterprise AI spending barely budges.
2. Regulation targets Nvidia’s dominance
Possible.
But also slow-moving and ineffective.
3. Competitors improve faster than expected
A credible risk.
But Nvidia’s lead is structural, not accidental.
4. Market rotation out of megacap tech
Short-term volatility? Yes.
Long-term damage? Unlikely.
5. Overreliance on hyperscalers
A concentration risk.
But hyperscalers represent both the demand and the distribution.
X. The Long-Term Case: Nvidia Is the Next Microsoft, Not the Next Cisco
People love comparing Nvidia to the dot-com era’s darlings.
They’re wrong.
Cisco rode a bubble.
Nvidia rides utility.
Cisco sold routers.
Nvidia sells the future.
Cisco sold hardware.
Nvidia sells infrastructure, software, and compute itself.
Cisco peaked.
Nvidia is still climbing its first mountain — with multiple mountains behind it.
XI. Price Targets? They Don’t Matter — The Trend Does
Analysts throw around price targets like:
-
$900
-
$1100
-
$1500
-
$2000
-
$2500
But these are just plot points in a decade-long arc.
The real story is:
Nvidia will keep compounding as long as AI demand compounds.
And AI demand is compounding exponentially.
XII. Final Conclusion: Nvidia Is Not Overpriced — Everything Else Is Under-Innovating
When a company:
-
grows faster
-
innovates faster
-
earns more
-
dominates more
-
builds moats wider
-
expands profitably
-
controls the ecosystem
…and is still cheaper than its Big Tech peers…
You don’t call that expensive.
You call that a gift.
Nvidia is now the second cheapest stock in the Mag 7 not because it’s fallen from grace, but because it has ascended so quickly that its valuation can’t keep up.
You can fight that.
Or you can understand it.
Some companies become great.
Some become legendary.
Some become infrastructure.
Nvidia became infrastructure.
And infrastructure isn’t a bubble.
It’s the world underneath the world.