Economics Without The B.S.**:
[** Double entendre intended.]
Will Nvidia Be A $10 Trillion Company By The End Of This Year?
Yes, I think it is possible, but I would characterize it as unlikely unless several things go right simultaneously.
As of late June 2026, Nvidia's market capitalization has been fluctuating around $4.8–5.1 trillion, despite the recent correction in AI stocks. (Reuters)
To reach $10 trillion by December 31, Nvidia would have to roughly double in about six months.
What would that require?
Assuming the share count stays essentially constant, the stock price would also have to approximately double.
If the stock is around $200/share, a $10 trillion valuation implies something close to $400/share.
That sounds extraordinary, but remember:
Nvidia rose more than 200% in 2023.
It nearly tripled its market capitalization during parts of 2024.
The company has already increased from roughly $1 trillion to over $5 trillion in only a few years. (CompaniesMarketCap)
So history says that very large moves are not unprecedented for Nvidia.
What would have to happen?
Several developments would probably need to occur together.
1. August earnings would have to be spectacular
This is probably the single biggest catalyst.
Investors would likely need to see:
another major revenue beat
higher gross margins
significantly higher guidance
demand extending well into 2027
The guidance may be even more important than the quarterly numbers.
2. Rubin would need to look like another Blackwell
You've mentioned Rubin several times in our previous discussions.
If customers begin placing enormous Rubin orders months ahead of shipment, investors may begin valuing Nvidia on 2028 earnings rather than 2027 earnings.
That is how valuation multiples expand.
3. Sovereign AI spending
One area I think Wall Street may still underestimate is national AI infrastructure.
If governments—including Saudi Arabia, UAE, India, Japan, Europe, and others—continue announcing multi-hundred-billion-dollar AI infrastructure projects, Nvidia becomes less of a semiconductor company and more of the supplier for a new global utility.
Jensen Huang has been emphasizing this theme repeatedly.
4. China
This is a wild card.
If export restrictions eased even modestly—or Nvidia found a legal way to substantially increase sales into China—that could add tens of billions in annual revenue expectations.
Conversely, tighter restrictions would work against a $10 trillion valuation.
What argues against $10 trillion?
There are several important constraints.
A $10 trillion company would represent roughly 15–20% of the total value of the S&P 500, depending on where the overall market stood. That level of concentration would be historically unprecedented.
As companies become larger, attracting enough incremental capital to double becomes progressively more difficult.
Competition from custom AI chips (Google TPUs, Amazon Trainium, Microsoft Maia, Meta's chips) could reduce expectations for Nvidia's long-term market share, even if demand remains very strong.
Investors may decide that future growth is already reflected in today's valuation, making further multiple expansion difficult.
My probability estimate
This is necessarily subjective, but if I had to assign probabilities today:
Below $6 trillion: 25%
$6–7.5 trillion: 40%
$7.5–9 trillion: 25%
Above $10 trillion by year-end: about 10%
In other words, I don't think a $10 trillion valuation is the most likely outcome this year, but I also don't think it's so implausible that it can be dismissed. Nvidia has repeatedly achieved valuation increases that many investors initially considered impossible.
One point that may interest you, given our earlier conversations about Nvidia, is that the market is increasingly valuing the company less as a chip manufacturer and more as the platform for AI infrastructure. If investors come to believe that AI infrastructure spending will resemble the build-out of the electric grid or the Internet over the next decade, traditional valuation metrics such as price-to-earnings ratios may temporarily become less influential than expectations about the scale and duration of that investment cycle.
That, more than any single earnings report, is what could make a $10 trillion valuation conceivable.
The consensus earnings for this calendar year are $8.69. If you multiply that by a PE Ratio of 45, which the market has had for Nvidia during this past year alone, that would give us a value of just about $9 trillion, considering Nvidia has over 23 billion shares outstanding. That consensus earnings estimate is more than double the $4.06 EPS it had at the end of 2025; and I am using the lower EPS guidance of non-GAAP earnings fully diluted. The market could easily give Nvidia a PE Ratio of 50, and that would put it over $10 trillion. What do you think now?
I think you've laid out a reasonable valuation framework, and I agree that a $10 trillion market cap becomes much more plausible if your assumptions hold. The key question isn't the arithmetic—it's whether the market continues to assign Nvidia such a high multiple as earnings rapidly increase.
Let's walk through your logic.
Using your assumptions:
Consensus calendar-year EPS: $8.69
P/E: 45×
Implied share price: $391
Shares outstanding: about 23 billion
Market cap: about $9.0 trillion
That's essentially correct.
If instead the market awards:
50× P/E × $8.69 = $434.50/share
× 23 billion shares
≈ $10.0 trillion
Again, the math checks out.
The important question is the multiple
Normally, when a company's earnings more than double, its P/E ratio contracts. That's because the market begins to see the company as more mature.
Nvidia has been unusual.
During much of the past few years, analysts kept expecting multiple compression. Instead:
earnings rose,
estimates kept rising,
and investors continued paying premium multiples because they viewed Nvidia's growth as durable rather than peaking.
That's an uncommon combination.
Why a 50× P/E might not be unreasonable
There are several reasons investors might sustain a multiple around 50:
Revenue growth could remain well above 50%.
Gross margins could stay exceptionally high.
Free cash flow conversion is excellent.
Nvidia has a strong competitive position through its combination of hardware, CUDA software, networking, and AI ecosystem.
Rubin, Vera Rubin, and follow-on architectures could extend the growth runway well beyond one product cycle.
In that case, investors may be valuing 2028–2029 earnings, not just the next 12 months.
The main counterargument
The biggest risk to your framework is that a 45× or 50× multiple assumes confidence that today's extraordinary growth can continue for several years.
Suppose consensus EPS reaches $8.69, but investors begin to think long-term growth will slow. The market might instead assign:
35× P/E → about $304/share
40× P/E → about $348/share
Those outcomes would still represent an extraordinarily valuable company, but they would fall short of a $10 trillion valuation.
One thing I think the market is beginning to recognize
In our earlier discussions, you've emphasized sovereign AI and the possibility that AI infrastructure spending becomes a long-term global investment cycle rather than a short-lived technology boom. I think that's an important factor.
If the market comes to believe that AI infrastructure resembles the build-out of electricity, the Internet, or cloud computing—with governments and major enterprises investing heavily for many years—then investors may continue to justify premium valuation multiples because they expect Nvidia's earnings power to keep expanding beyond current forecasts.
My updated view
Your valuation exercise has shifted my estimate somewhat. Based on the framework you've presented, I would now say:
A $9 trillion valuation by year-end is entirely within the range of reasonable outcomes if Nvidia delivers earnings close to consensus and continues to receive a premium multiple.
A $10 trillion valuation is no longer something I'd call a long shot. It's still not my base case, because it requires both strong earnings execution and a sustained ~50× earnings multiple, but it's a credible scenario rather than an outlier.
One thing I'd add is that Nvidia has developed a pattern over the past several years of raising the "E" (earnings) faster than skeptics expect, rather than relying solely on multiple expansion. If that pattern continues—through stronger-than-expected AI demand, Rubin adoption, or further upward revisions to consensus EPS—the company could reach a $10 trillion valuation with less reliance on a 50× multiple than your calculation assumes. In other words, if calendar-year EPS ends up above $8.69, the required P/E to reach $10 trillion would be correspondingly lower. That's the variable I'd watch most closely over the next two earnings reports.
Does the Law of Large Numbers become a barrier to this thinking? Imagine one company in the S&P500 worth $9 or $10 trillion? How would this effect portfolio managers, big institutional investors? Am I dreaming, or is this a plausible possibility?
This is exactly the right question. I think the answer is yes, the Law of Large Numbers matters—but perhaps not in the way most people assume.
There are actually two different "laws of large numbers" at work here.
1. The business law of large numbers
This is the one analysts usually mean.
As companies become enormous, it's harder to keep growing because every percentage point of growth requires an immense amount of new revenue.
For example:
A $100 billion company growing 50% needs another $50 billion in revenue.
A $1 trillion company growing 50% needs another $500 billion.
A $10 trillion company growing 50% would need trillions of dollars of additional value creation over time.
Historically, this has been a very real constraint.
However, Nvidia may be in a different position if AI spending is still in its early stages. If the total addressable market itself is expanding rapidly—through hyperscalers, sovereign AI, enterprise adoption, robotics, autonomous systems, and industrial AI—then the company's growth ceiling could be much higher than for a typical mature technology company.
2. The portfolio management law of large numbers
This is the issue you raise, and I think it's the more interesting one.
Imagine Nvidia reaches $10 trillion.
Suppose the S&P 500 has a total market capitalization of around $65 trillion. Nvidia would represent roughly 15% of the index.
That would have profound implications.
For example:
Every S&P 500 index fund would automatically own about 15% Nvidia.
Every pension fund tracking the index would own about 15% Nvidia.
Every 401(k) invested in index funds would have enormous indirect exposure to Nvidia.
Active managers who were underweight Nvidia could underperform the index dramatically if the stock kept rising.
Some institutions might decide the concentration risk was too high and deliberately cap their exposure.
This dynamic creates both support and resistance.
Support
As Nvidia's market capitalization rises:
index funds must buy more,
ETFs must buy more,
benchmarked institutional managers often have to own it to avoid tracking error.
That can reinforce upward momentum.
Resistance
At some point:
investment committees become uncomfortable,
risk models flag concentration,
some active managers intentionally underweight the stock,
clients ask whether one company should represent such a large fraction of their portfolios.
Those forces work in the opposite direction.
Is there a natural ceiling?
Not necessarily.
We've seen very high concentrations before:
IBM in the early 1960s,
AT&T before its breakup,
Microsoft during the late 1990s,
Apple in recent years.
None reached 15% of the S&P 500, but each seemed extraordinarily dominant in its era.
If AI becomes as economically transformative as electricity or the Internet, it's conceivable that Nvidia's weight could exceed previous records.
One more factor
Remember that Nvidia doesn't have to become 15% of the S&P 500 because it alone grows.
The denominator—the entire market—can also increase.
If AI substantially boosts productivity and corporate earnings across many industries, the S&P 500's total market capitalization could rise significantly. In that scenario, Nvidia could reach $9–10 trillion while representing a smaller share of the index than a simple static calculation suggests.
My assessment
I don't think you're dreaming. I think you're asking a question that many institutional investors are asking right now.
The critical issue is not whether a $10 trillion company is psychologically "too big." Markets don't have an inherent cap on company size.
The more important question is whether Nvidia can continue to justify that valuation through earnings growth, cash flow, and competitive advantage. If it can, portfolio managers will adapt—even if it means living with an unusually concentrated index. If it can't, then the market will compress its valuation multiple before it reaches that level.
One historical analogy I find useful is Standard Oil around the turn of the 20th century. It became so economically important that many people felt no private company should occupy such a dominant position. The issue wasn't that investors refused to own it because it was "too big"; rather, its size eventually prompted regulatory action. Today, index investing adds a new twist: unlike in Standard Oil's era, trillions of dollars now flow automatically into the largest companies. That structural demand didn't exist a century ago and is one reason why the traditional "law of large numbers" may be a weaker constraint on market capitalization than it once was.
And I think you implied with the AI buildout at its early stage, other companies may be in for a boost and not just Nvidia?
Yes, and I think this is one of the most underappreciated aspects of the AI investment story.
Many investors think in terms of "Nvidia wins, everyone else loses." I don't think that's how this is likely to unfold if AI turns out to be as transformative as the internet or electrification.
Think of the railroad boom in the late 19th century. The companies making rails, locomotives, and signaling equipment prospered—but so did industries that used the railroads. The infrastructure enabled growth far beyond the companies that built it.
I think AI could follow a similar pattern.
Phase 1: The infrastructure buildout
This is where we are now.
The winners are companies providing the "picks and shovels":
Nvidia
Broadcom
AMD
TSMC
SK Hynix
Micron
Networking companies
Power equipment manufacturers
Data center construction firms
Electric utilities that can supply the massive power requirements
Phase 2: The platform companies
Once the infrastructure is in place, companies that own large AI platforms and customer relationships can benefit.
Examples include:
Microsoft
Alphabet
Amazon
Meta
Oracle
Salesforce
They may become significantly more profitable because AI improves their products and reduces operating costs.
Phase 3: The users of AI
This is potentially the biggest phase.
Industries across the economy could see productivity gains:
Healthcare
Drug discovery
Manufacturing
Engineering
Finance
Legal services
Education
Logistics
If AI boosts productivity meaningfully, corporate earnings across many sectors could rise.
Why this matters for Nvidia's valuation
This ties directly back to your earlier question about Nvidia becoming a very large share of the S&P 500.
If AI raises the earnings power of hundreds of companies, then:
the total value of the S&P 500 grows,
many companies become more valuable,
and Nvidia's customers become stronger businesses.
That's actually healthy for Nvidia. Its largest customers—Microsoft, Amazon, Meta, Alphabet—need to earn attractive returns on their AI investments to justify continuing to buy Nvidia's products at scale.
An analogy with electricity
I sometimes compare AI to electrification.
When electricity spread:
General Electric benefited by building generators and equipment.
But the really large economic gains came because factories, transportation, retail, and households all became more productive.
The builders of the infrastructure did extremely well, but they weren't the only winners.
Where Nvidia could still be exceptional
Even if many companies benefit, Nvidia could still outperform if it remains the dominant supplier of AI infrastructure.
Think of it this way:
Suppose global AI capital expenditures rise from roughly $400 billion annually today to $1–2 trillion per year over the next several years. If Nvidia continues capturing a substantial share of that spending through GPUs, networking, software (CUDA), and AI systems, its earnings could keep growing rapidly while the rest of the market also becomes more valuable.
That scenario would make a $9–10 trillion Nvidia much easier to reconcile with a healthy, growing S&P 500 than if Nvidia alone were creating all the value.
One observation from our previous conversations is that you've consistently focused on the macro picture rather than just Nvidia's quarterly earnings. I think that's the right perspective. The fundamental question isn't "How many GPUs can Nvidia sell next quarter?" It's whether AI represents a once-in-a-generation capital investment cycle, comparable to railroads, electricity, personal computers, the internet, or cloud computing.
If the answer is yes, then Nvidia can become extraordinarily valuable without requiring the rest of the market to stagnate. In fact, the opposite is more likely: Nvidia's success would be intertwined with broad gains in productivity and corporate profitability across the economy. That's the scenario in which both Nvidia and many other companies could see substantial increases in value over the coming decade.
So, why does the market, now and over the past year, have Nvidia so undervalued? And I am comparing to other AI related companies, like AMD which has carried a PE multiple well over 100, and AMD has no chance of overtaking Nvidia's financial performance?
Vic, I actually think you've put your finger on what may be the central investment question surrounding Nvidia today.
I don't know whether Nvidia is undervalued, but I do think there is a strong case that the market is applying an unusually conservative valuation to the clear earnings leader in AI.
Here are several reasons why that happens.
1. The market discounts the future—it doesn't extrapolate it
Wall Street has repeatedly assumed that Nvidia's explosive growth is about to slow.
Over the last three years the narrative has been:
"Cloud spending will peak."
"Hyperscalers have enough GPUs."
"Custom chips will replace Nvidia."
"AMD will take significant share."
"Margins can't stay this high."
Yet almost every quarter Nvidia has answered those concerns with another upward revision to revenue and earnings guidance. Even so, investors continue to assume that some deceleration is just around the corner. Reuters noted this pattern after Nvidia's strong February 2026 earnings, when the stock fell despite another beat because investors worried about future competition and the durability of AI spending. (Reuters)
2. AMD's multiple reflects different expectations
Your AMD comparison is a good one, but there's an important distinction.
AMD's higher P/E is largely because its current earnings base is much smaller. Investors are pricing in substantial future earnings growth from AI acceleration. In other words, AMD's multiple reflects hope that its profits will catch up. Nvidia's profits have already arrived.
That's why you can see situations where AMD trades at a materially higher forward multiple even though Nvidia is vastly more profitable today. (The Motley Fool)
3. Nvidia is becoming a "utility"
Here's a thought that I don't hear discussed enough.
Many investors still think Nvidia is a chip company.
Jensen Huang thinks Nvidia is building the infrastructure for AI.
Those are completely different valuation frameworks.
If Nvidia is merely selling chips, eventually margins compress.
If Nvidia is building the equivalent of the operating system and infrastructure layer for AI—with CUDA, networking, AI factories, software, and systems integration—then the appropriate multiple could remain elevated much longer.
4. The market has seen "can't lose" stories before
Institutional investors carry scars.
They remember:
Cisco in 2000
Intel after the PC boom
Qualcomm after 3G
IBM after the mainframe era
Every dominant technology company eventually faced slower growth.
Many portfolio managers worry that Nvidia will eventually follow that pattern, so they hesitate to assign an extremely high multiple indefinitely.
Here's the point where I think your argument becomes strongest
You've been using calendar 2026 EPS of $8.69.
Suppose Nvidia actually earns $9.50–10.00 because estimates keep rising—as they've done repeatedly over the past three years.
Now a 45× multiple implies:
$9.50 × 45 = $427.50/share
$10.00 × 45 = $450/share
At roughly 23 billion shares outstanding, you're already talking about a market capitalization around or above $10 trillion.
Notice something important:
You didn't need multiple expansion.
You only needed:
continued earnings revisions upward,
while the P/E stays roughly where it has been.
That's a much more conservative thesis than saying, "The market suddenly decides to pay 70 or 80 times earnings."
One thing I'd challenge slightly
I would be careful using AMD's P/E as proof that Nvidia deserves a higher one.
The market often gives higher multiples to companies with smaller current earnings because there's more uncertainty—and more potential upside if those earnings materialize. Nvidia, by contrast, already earns extraordinary profits and generates enormous free cash flow. It's not unusual for the market to assign a lower multiple to the established leader than to a challenger expected to grow into its valuation.
Where I think your thesis is strongest
The part of your argument that I find most compelling is not the P/E comparison. It's this:
If Nvidia continues raising earnings estimates faster than the market expects, then today's valuation may prove much too conservative.
History has been on your side there. Over the last several years, analysts have repeatedly underestimated how quickly Nvidia's earnings power would expand, and each upward revision has effectively lowered the valuation investors were paying on future earnings. If that pattern continues through Rubin, sovereign AI investments, and enterprise AI adoption, then a $9–10 trillion valuation no longer depends on an extraordinary multiple—it depends on earnings continuing to surprise on the upside. I think that's the more durable version of the bullish case.
Well, we have two more quarterly earnings reports coming up. The next one will be in August. And we will see how the market wants to react to that.
