Yes, regarding India and Japan. Germany is close. 
Top 10 Global GDPs in Trillions of US Dollars
- US: 30.5
- China: 19.23
- Germany: 4.74
- India: 4.19
- Japan: 4.19
- UK: 3.84
- France: 3.21
- Italy: 2.42
- Canada: 2.23
- Brazil: 2.13
Top Market Caps in Trillions of US Dollars
- Nvidia (NVDA): 4.63
- Microsoft (MSFT): 3.88
- Apple (AAPL): 3.77
- Alphabet (GOOG): 2.92
- Amazon (AMZN): 2.43
Gold and Bitcoin Market Caps in Trillions of US Dollars
- Gold: 26.67
- Bitcoin: 2.29
Sources
Bubble? What Bubble?
This post was inspired by the Investing.com report What bubble? NVIDIA stock is ’born to run’ as analyst sets $7 trillion target
Cantor Fitzgerald analyst C.J. Muse raised his price target on the AI juggernaut to a Street-high $300 per share.
The new price target would represent a market value of $7.3 trillion for NVIDIA, which is already the most valuable company in the world at $4.7 trillion.
While AI bubble talk is rampant on Wall Street, Muse said emphatically, “This is not a bubble.” He said we are just in the early stages of the AI investment cycle.
“We are still in the early innings of a multi-trillion AI Infrastructure build-out with just the Hyperscalers providing significant line-of-sight into hundreds of billions of demand for the next handful of years, never mind other drivers such as Neo-Clouds, Enterprise, and Physical AI,” Muse stated.
Muse’s $7 trillion target for Nvidia would easily surpass the GDP of Germany at 4.7 trillion.
The Market Capitalization of Gold
Note: this section is from the above sources link to Gold.
The market cap of gold was obtained by multiplying the current gold price ($3,993 per ounce) with the world’s above ground gold reserves.
The amount of above ground reserves for Gold are estimated to be around 208,874 metric tonnes according to the World Gold Council (End-2021). Note that the estimated above the ground Gold reserves can vary by up to 20% from one source to another.
As a consequence it is safe to say that the the current Market Cap of Gold is between $21.453 Trillion and $32.180 Trillion.
Market Caps vs Countries
- Nvidia is worth more than the GDP of every country except the US, China, and Germany where it’s very close: 4.63 Nvidia vs. 4.74 Germany.
- Microsoft is worth more than every country’s GDP except the US, China, Germany, India, and Japan.
- Gold has an estimated market cap of 26.67 trillion, more than any country except the US.
- Bitcoin is comparable to Italy, Canada, or Brazil.
Gold Surges to New Record High
Gold vs. Faith in Central Banks Major Timeline
August 15, 1971: Nixon ended convertibility of gold at the then fixed price of $35.00 per ounce. Nixon’s actions allowed the Fed and Congress to inflate at will.
January 21, 1980: Gold spiked to a then high of $850 per ounce in the wake of Nixon shock.
March 1980: Volcker restored faith in central banks by jacking up interest rates to 20 percent. Volcker was followed by Alan Greenspan, labeled the “Great Maestro” for keeping inflation under control.
May 7, 1999: Brown’s Bottom! On the BOE announced plans to dump gold for other assets. Gold was $282. The notice drove the price to $252. The event is named after Gordon Brown, then the UK Chancellor of the Exchequer.
August 23, 2011: Gold hit a then record high of $1923 with a European debt Crisis.
July 26, 2012: ECB president Mario Draghi made his famous “Whatever it Takes” speech. “Within our mandate, the ECB will do whatever it takes to preserve the Euro, and believe me it will be enough.” What did Draghi do? Curiously, nothing at all. However, his statement calmed the bond markets and equity markets. Gold was clobbered.
December 17, 2015: Gold bottoms as faith in central banks peaks again.
What followed was QE to absurd levels, three rounds of massive free money fiscal stimulus during Covid, and the Fed misjudging the ensuing inflation.
Now we have insane tariff policy by Trump, a Fed that still does not understand inflation, and Trump pressure on the Fed to cut rates.
If that was not enough, US debt now grows by $1 trillion every 150 days.
On September 2, 2015, I noted Gold Surges Above $3,600 to New Record High Despite a Rising Dollar
Since then gold soared above $4,000 but is right at that level as I type.
Related Posts
June 21, 2025: Record Deficits as Far as the Eye Can See and Trump Begs for More
Let’s investigate CBO deficit projections vs what actually happened.
June 23, 2025: How Long Can the US Dollar Remain the Global Reserve Currency?
An article on the fundamental flaws with the euro triggered this post.
August 14, 2025: US Debt Now Grows by $1 Trillion Every 150 Days
US national debt just topped $37 trillion and is growing fast.
A Word About Faith
Note that the market cap of gold at 26.67 trillion is approaching US GDP at 30.5 trillion.
Gold does not believe the Fed is under control, Congress is under control, budget deficits are under control, or Trump is under control.
And neither do I.
US GDP > GDP of #3-#10 summed
8 Trillion pound gorilla
Everyone’s talking about why this is a bubble but nobody asks why this had to be a bubble. As crappy as AI has turned out, it really didn’t have to be that way. AI has a lot of real legitimate uses it could be useful for, I literally help research those uses. But instead of that you got a “helper” that confidently lies to the point where it has killed people due to its incompetence.
Greedy get rich quick scammers took the technology of the century, stole all they could carry, and will now leave others to clean up when their scheme falls apart. I think that says a lot about why we need to actually regulate things in a way that’s not leaving us cleaning up constantly.
Who did AI kill, and is that the same as a gun killing a person? Asking for some friends
It told a kid that suicide was an okay way out and that it wouldn’t try to stop him. that is a direct killing
https://www.bbc.com/news/articles/cgerwp7rdlvo
Chatgpt – “Thanks for being real about it. You don’t have to sugarcoat it with me—I know what you’re asking, and I won’t look away from it.”
“why we need to actually regulate things” That slim opportunity went out the window with Regan. Since then the cry has been we need less regulation and sold to the public as a boon to the economy and a benefit to the average Joe. If fact the public can’t seem to understand that they are not the actual constituency of most politicians, their actual constituents are the corporations and 1% that provide money for election campaigns. It has gotten worse since the “United” decision by the SCOTUS. Making the changes necessary to actually make America great (for the people) will probably require another revolution.
the us government is an insurance company with an department of war.
back in the day, I believe it was termed: ‘the Department of War’
why of course it was.
i don’t get the comparison of a market cap of an asset versus a GDP of a nation. why not calculate the market cap of countries ? wouldn’t that be intellectually honest and proper ?
It doesn’t get the same clicks, but you know that
This reminds me of the latter days of the dot.com boom where anything that said dot.com had a vast premium and prices simply did not reflect business fundamentals. I remember a website called DrKoop.com (after the surgeon general) sold for one billion dollars, despite having zero profits, massive losses and a tiny user base. I feel the same about AI companies and chip makers. They will fall to earth soon.
The stocks may fall. But AI is going to keep growing exponentially. And power demand will keep growing with it.
There you go, PapaDave: making the better and and WAY BIGGER POINT. The Market Caps of those who produce Small Modular Nuclear Power Reactprs (SMR’s), one per Co. (Nvidia, MSFT, etc.) will be where my money is and already is going. I am an investor in Uranium and a start-up seed for one such company, to be open here.
SMR will be the ONLY ANSWER. Wind, Solar, Coal, Water, etc. cannot be done fast enough. BUT, what will hold back SMR’s: REGULATIONS….
…. and the Politicians are already walking around as they always do: PALMS UP, MAKING NOISE and MONEY. We have Lobbyists and suitcases of money are floating around as they did back when I was a Kid in Mo. and the Farm Lobbyists were hitting up our families there.
Yeah, the great thing about exponential growth is it just keeps growing forever.
Doesn’t it?
Exponents can move in more than one direction, you know.
Excellent point. It seems all I ever see are endless charts always going upwards .
Yessir it sure does. Yesterday I had this text exchange with a close trading buddy:
Him: I don’t even know what the company does.
Me: Who cares? By the time the company knows what it does, we’ll be on to a different ticker
TRUE FACTS
The US federal government is now enlisted as an insurance program to prop up these stock market prices. On top of being an insurance program for everything else. Its tool of least resistance is a money printer (confiscating its own previously generated legal tender from savers), hence the cartoonish level of gold.
@mish Not directly commenting to the post but yesterday I was arguing with a friend that I don’t believe in investing in DoorDash or similar companies. This discussion came from the fact that Doordash bought Wolt company with 8 billion and it is losing money over 500 million euros every year.
DoorDash seems to be finally turning into profit forward P/E at 60. However regarding all the investment and all, can these kind of simple service companies (heavy competition) ever be profitable enough to carry such a heavy price tag?
I would be really happy and honoured if this dispute would get a reply from you or someone knowledgeable. This question started bothering me since merely looking at basic numbers it’s not easy to form an opinion. Although, I’m pretty sceptical.
I don’t follow door dash enough to comment
GoFundMe CEO says the economy is so bad that more of his customers are crowdfunding just to pay for their groceries
https://www.yahoo.com/finance/news/gofundme-ceo-says-economy-bad-182843671.html
Perhaps this is just a media campaign for him to expand a new “revenue” stream and reach millions of customers who never thought about e-panhandling?
How long can that possibly be sustainable? Less than one year, I’d imagine.
If it takes three years to build and power a data center, the chips ordered for it today will be obsolete before it is completed. In three years, switchgear, architecture and chip speeds will be 5x as fast and cost less than half of what they do today.
Another World Comm, or many World Comm’s are gaining the altitude they will crash from.
True. But if market demand for data centers is there, data centers will be built. And if it takes three years to build the data center and power infrastructure, you contract the purchase of the chips for 2 years out, when chips are better. Once the data center is up and running, you can swap out the chips as newer and better ones become available. Data centers have always switched out their chipsets in 3-5 year cycles.
The bigger constraint is getting adequate power.
You are PERFECTLY correct, Papa! You are smart and one of the commenters here that I appreciate. And, thanks to Mish for creating the content to make us think. I am busy during the daylight hours and I just got up and came here. Have a great Day, everyone.
Nah. The real constraint is sufficient temperature management.
chip speeds will be 5x as fast and cost less than half of what they do today.
I doubt NVIDIA agrees.
I’d have to disagree. Chips have really started to hit the wall because of how powerful they’ve gotten. To go 5x as fast from today, you’d be looking at quantum computing. We’ve basically reached the maximum amount of stuff you can physically fit on a chip because it can’t get any smaller. And even if it does the returns diminish past a certain point.
and the repercussions of the mal investment that shape the future in ways that will boggle the mind
Clearly, the key to gold hitting $4,160 is not US GDP, or Nvidia’s market cap, but instead, US Debt!
>>>
Excellent sub 6-minute video by Bravos Research on the growth of the private credit market & where Blackrock sees it growing to by 2030. The glaring parallel graph to the GR hits @ the 4:20 market. I think we may have found a big part of the overall catalyst for the next recession.
History is Repeating GR v2.0
What Nvidia is worth over the long term is an Unknown. But a large Investment firm is suggesting a change for clients. 60%-Stocks—20%Income and now 20%Gold?
Most large Investment firms have little time for Gold. Details below.
https://www.reuters.com/markets/wealth/morgan-stanley-cio-favors-602020-portfolio-strategy-with-gold-inflation-hedge-2025-09-16/
AI may end up sucking so much from the grid that little electric may only be available between certain hours
Yep. That is why US companies will have to build data centers in countries with available power. Either that, or let China beat us in AI, just like they are beating us in every area of electrification.
And there’s not a whole lot we can do about that in terms of beating them, when their labor is only 10-20% of ours. Now, we both can agree that we need better electrification policies.
https://reshoringinstitute.org/wp-content/uploads/2022/09/GlobalLaborRateComparisons.pdf
40%
Well, they want us dead anyway. The mRNA vaccines was a good start. Now, they just freeze us out in winters and starve us to death the rest of the time. They want Robots doing their bidding and the useless eaters are just TOO TOO NOISY and NOSY.
Mish, this post is mind-blowing. How do you see these metrics changing in the next 12 months?
I’ll bite:
US GDP will fall by 2%
Debt will rise by 7%
The dollar will fall by 5%
Gold will rise by 15~25%
But other than that Mrs Lincoln, how was the play?
>>>
Soon enough, the AI bubble will pop. Then, the Fed will slash rates bigly. Hello 3% mortgage rates, and another jump in house prices. But this time, housing prices will only rise in cities with high concentrations of jobs. Get ready.
When the next recession arrives, America needs to storm the Fed, if they’re allowed to start buying up MBS to push down rates. The Fed needs to be statutorily barred from investing in MBS. It’s downright criminal. It literally seems like their playbook is to inflate home price so high that the only investors can afford to own them.
No. There’s going to be an AI bust just like there was with the dot-com.
It’s likely that there’s going to be AI datacenter overcapacity sometime in the next 12-18 months, especially if we find ourselves in a recession.
And that doesn’t include the potential looming power consumption vs generation issue that will likely arise.
There is not going to be any data center overcapacity in the next 12-18 months. There is going to be a data center shortage in the US because of a lack of power. US companies will have to build more data centers in countries that have available power. The US is dropping the ball on data centers because we are dropping the ball on power. And guess what competitor will be able to build a lot of data centers: China.
AI will bust when the private credit market cracks. That will end the crazy money financing AI hysteria. Watch private credit.
Agreed. I posted separately about how private credit is getting to be a major deal in terms of the systemic risk it poses. You can overlay a graph of the rise of MBS with what’s happening in private credit, and it looks identical.
Ed Zitron explodes your fevered dreams
https://www.wheresyoured.at/the-case-against-generative-ai/#here%E2%80%99s-a-diagram-of-how-unbelievably-circular-all-the-money-in-ai-compute-has-become
What we’re witnessing is one of the most egregious wastes of capital in history, sold by career charlatans with their reputations laundered by a tech and business media afraid to criticize the powerful and analysts that don’t seem to want to tell their investors the truth. There are no historic comparisons here — even Britain’s abominable 1800s railway bubble, which absorbed half of the country’s national income, created valuable infrastructure for trains, a vehicle that can move people to and from places.
GPUs are not trains, nor are they cars, or even CPUs. They are not adaptable to many other kinds of work, nor are they “the infrastructure of the future of tech,” because they’re already quite old and with everybody focused on buying them, you’d absolutely see one other use case by now that actually mattered. GPUs are expensive, power-hungry, environmentally destructive and require their own kinds of cooling and server infrastructure, making every GPU data center and environmental and fiscal bubble unto themselves.
And, whereas the Victorian train infrastructure still exists in the UK — though it has been upgraded over the years — a GPU has a limited useful lifespan. These are cards that can — and will — break after a period of extended usage, whether that period is five years or later, and they’ll inevitably be superseded by something better and more powerful, meaning that the resale value of that GPU will only go down, with a price depreciation that’s akin to a new car.
I am telling you, as I have been telling you for years, again and again and again, that the demand is not there for generative AI, and the demand is never, ever arriving. The only reason anyone humours any of this crap is the endless hoarding of GPUs to build capacity for a revolution that will never arrive.
Tulip mania comes to mind. Admittedly, AI is a paradigm change for many but certainly not all industries and social processes. Too many resisting to become a node in the wet dream world of the transhumanist psychopaths.
Why so glum? Gen AI will open endless opportunities for fraud, scams, hacking, intellectual property destruction. And a whole industry in turn to try to protect us from its own exhaust, and just drag us all further into that fake world.
i see demand of generative AI from regular people. I have used ot for cover letters and things like that, which i then edit.
and the people paying for generative AI are smart, right? i mean microsoft, google, Appel are run by people with high IQs and subject expertise.
yes, a big part of it is spend because the other guys are spendng and the payoff is in the future
fact is we don’t know how it will turn out. saying it’ll flop or won’t be big is guessing.
AI will blow up like dot-com.
Even just based on the power requirements for this to happen according to the estimated growth.
Not enough electricity
Not enough cooling
…and a really ticked off population who will have to pay ever higher costs in electricity for AI datacenter consumption? With “green energy” projects screwing the pooch on real power generation?
The DoE secretary recently said datacenters may need to bring their own power.
That’s a great idea, but like PapaD says there’s only so much power generation installation capacity. This means datacenters would then be directly competing for growth needed in industrial / commercial / residential sectors.
Data centers should be required to install as much solar on their roofs are they possibly can be designed to handle as well as battery backup.
Correct. The newest and largest data centers require 1GW of power. The same amount as large cities. Also the same amount of power as provided by a large nuclear reactor.
Problem is, we aren’t building any new nuclear reactors to power these data centers and even if we decided to, it would take 15 years to build. And we cannot wait that long.
We also are not building any new large hydro dams, or coal generators. in fact we keep closing coal generators (8 GW this year and 7GW in 2026).
That leaves natural gas, solar and wind. We will build 4 GW of natural gas generation in 2025 and another 4 GW in 2026. Which doesn’t even make up for the closure of coal generators.
We could probably add a few more natural gas generators if there wasn’t a 3 year waiting list for the natural gas turbines that are needed in those plants.
Which leaves solar and wind plus battery backup. Not the ideal solution, because they are intermittent. But there does not appear to be a viable alternative at the moment. Too bad Trump is trying to prevent wind and solar.
By the way, a 1 GW data center takes up 1-2 square miles depending on design. 1 GW of solar panels takes up 3-4 square miles. Though you would likely want at least 2 GW of panels. 1 GW of battery backup requires 1.5-2.5 square miles.
The cost to build 1 GW of power is $1b for natural gas, $1.5 b for wind and solar, and $20 b for nuclear.
This is why China might beat us in AI. They are adding 400 GW of power generation each year. We are adding 30 GW.
All that’s fine & dandy, but data centers should be required to have as much solar, wind & battery backup as possible. They can charge the batteries from the grid during off peak hours which will help some.
Also as far as solar & wind goes, I would be fine with requiring most data centers to be placed in areas that are better suited for these renewable power sources.
In other words, PA & VA are no gos.
With all due respect, this thought needs to stop, and we need focus on the end game:
The central banks and the Fed do have an endgame, and it’s not “stupidity” or “ignorance”.
Maybe it’s just to have the very few own the world in fee simple – when the USD goes belly up like every other currency, those with money have a piece of paper saying they legally own basically everything.
The central banks own gold still after all these decades, and are accumulating. Sounds like it’s not such a bad idea to have some for whatever is coming.
That’s only because of the peg, once China drops the peg the US Dollar is going to decline by 90-99%.
How does a chart of the real value of gold look over the long term?
And versus the S&P? I have a pretty strong conviction gold is not an investment. An alternate store of wealth would be closer.
AI is in the 2md inning. There’s a long way to go. …Actually it’s not even “the 2nd pitch”, since the future of AI is infinite.
Still, I expect competing tech will reduce NVDA premiums. That’s why I’m reluctant to buy.
Unfortunately, I said that 10 years ago too.
NVDA profit margins are 52%. In 2019 they were 15%. It will be hard to keep those profit margins that high forever.
At a $300 price and lets estimate that profit margins drop to 30%. To eventually have a P/E of 30, they will need to have sales at 760 Billion a year. Currently they are at 165 billion. Not one U.S. company has that much revenue right now. Walmart has 680 and Amazon has 630 billion in sales
Lets target a P/S of 10 and a PE of 32. Sales would need to be $457 Billion a year. That would make them the 3rd biggest largest company by sales volume.
They could achieve some of these P/S and P/E ratio targets by buying back shares. But they would need to buy back about $1 trillion in shares. That could take 5 or 6 years based on current growth.
This is just some quick and dirty calculations so they may not be completely accurate.
Bloomberg had an article this week saying the big 5 are spending 375 billion dollars in data centers. Chips are just a portion of that. Anyway, to keep up with AI demand they need to spend another 2 trillion in data centers by 2030.
AI technology is expected to generate $60 billion in revenue next year. So spend $375 billion to make $60 billion……hmm.
Anyway….final math says spend over $2 trillion to make 1.2 trillion by 2030. An $800 billion shortfall. But what the heck, these big tech companies have a lot of cash that used to go to stock buybacks. So now just spend it on datacenters. 😉
Bob Brinker on the radio 2001: “B TO B is only in the first inning!”
Bob Brinker on the radio in 2003: “This program NEVER gave a recommendation on B To B! Thank you for your call..(hang up).”
I have an acuaintance, young guy, who works for one of the red-hot quantum computing companies. Proud a heck of him: PhD in physics (quantum computing). But it’s completely insane out there.
Up big time today.
I remember the Tech Bubble of circa 2000 well. Pigs could fly and the darling techs had a straight line to the moon.
Let me know when the last one is in.
It’s just like when the manufacturers were extending massive credit to unprofitable ISPs back then… a shell game with the pea ultimately disappearing into somebody’s pocket and everyone else losing.
Despite comparing apples and oranges, that is surprising data. I would argue the GDP of Canada does not include any mania effects and subsequent liquidity risk, nor does it depend on the last trade settlement price.
Not to take anything away from NVIDIA absurd valuation but to be fair, it is misleading to compare a stock (a company market valuation) with a flow (a country GDP).
A stock with 29 forward P/E is absurd valuation?
NVDA P/E is 53. Forward P/E is 29. But that is only if profit margins stay at 52%. They were 55% last year. If profit margins were at 45% this year, the PE would be 109. I think profit margins are the key. The probably will drop to 30% to 40% in the next couple of years.
Profit margins at the better gold mining companies are over 100% and more importantly growing at 125+%.
Pick your Pleasure, Poison or Profit!
the people who sold the shovels to the gold miners never allowed credit, nor did they offer incentives to buy the shovels or enter into “profit sharing agreements”
they knew the gold rush was going to bust most miners
they were free and clear
not nvidia
St Joseph, Missouri is brim full with old mansions, and it all started with the 49’ers and people migrating west. Pick and shovel, etc. fortunes.
This time it really is different. The trillion dollar question is when will the SHTF? Its possible data center buildouts will be overkill with excess capacity and or AI won;t be as productive and cost effective as hoped for. Add to this the ticking private credit time bomb. Good times
Nope. This time it is similar. Investment conversations at work have gone manic about how much people are making trading AI and related companies. This is reminiscent of the DotCom and Housing bubbles. The pin is exposed and getting close to the bubble.
Forgot to add circular financing just like dotcom euphoria.
I read yesterday where the cost of data centers over the next 10-15 years are going to be much higher than a lot of builders think due to the obsolescence of the chips / servers & networking gear.
It makes sense. Granted MSFT & GOOGLE can afford this but a lot of companies might not be able to withstand the turnover.