Complacency

 The music is playing and we ought to dance. I think this is what Howard Marks' memo is saying.

I was rewatching Margin Call yesterday: fabulous drama. 
And what struck me was the dichotomy between ex-post and ex-ante.

Back in 2006-7; the overwhelming majority was aware that 'this is too easy' and that any mention of 'risk' was rubbished away. There was genuine complacency.

Just like 1999, where anybody who got out before the bubble burst, felt left out, dumb and cheated. 

That, even in Margin Call - once the 'company' was done selling its assets, if the market recovered and continued chugging along: everyone felt like a fool.

"May be there is some oil to be found in hell".

Ex-post however, everyone sees it as alarmingly obvious; "if I was there then, I would have surely figured it out and gone into cash or sold short..."

And yet, here we are. 

Today:

"I own great companies. So what if some of them are at 40x PE"
"Well, yes. But I have 4% in cash, and look at these bargains in the other corner of the market"

As on Nov 25: Mag 7 total + Oracle + Broadcom : All 9 combined: $23.692T ( of 67.5 T US total)

As on 2019-12-31: 5,448.13 (of 33.9 T US total)

The magnitude of this is alarming. 


From the movie:

"It's just money; it's made up. Pieces of paper with pictures on it so we don't have to kill each other just to get something to eat. It's not wrong. And it's certainly no different today than it's ever been.

1637, 1797, 1819, 37, 57, 84, 1901, 07, 29, 1937, 1974, 1987—Jesus, didn't that **** me up good—92, 97, 2000 and whatever we want to call this (2008).

It's all just the same thing over and over; we can't help ourselves. And you and I can't control it, or stop it, or even slow it, or even ever-so-slightly alter it. We just react."

So, 2020 got a COVID crash; but the recovery was mad quick; there was a bit where the end of ZIRP took down a few big banks in the US, and nobody really cared. 

My point being: 

The attitude towards risk is - more is good.

There is complacency towards the downside.

Freedom's Forge - AI

I was once asked, whats the point of reading history books when investing is all about the future?

Freedom's Forge was a great book on how the US marshalled its resources to supply materiel (as it is called?) for Europe, USSR and for itself to be a great force to beat the Axis powers.

And 2025 reminds me of that.

The world is not at war right now, but the AI wave has the US pitted against another great economy which is a lot more 'closed'. And the US cant explicitly fund the massive buildout, whereas China and its top-heavy machinery can. 

For example, China can marshall resources to build out a $200 B network of computing and data centers / AI factories. But the US has to rely on free market forces?

So, the current wave of 'circular deals' where OpenAI gets funding from NVIDIA and Msft, and now Anthropic has a big deal, etc. all point to an unsaid commandment from the US Govt.

That 'we have to win this'. Reminds me of the Space Race and the Nuclear race; if we win this AI race, we get to be the dominant economy, and dominant geopolitical force.

If it sounds like a duck and walks like a duck... 

I guess thats what I am alluding to; that these large tech companies have effectively zaibatsu-ised and interlinked to make sure that nobody goes bankrupt, and that innovation can happen.

This might be one of the greatest leaps of innovation that the world has seen.   

Avalanche

 A fragile system.

It looks like thats where we are.

Pvt credit using insurance subsidiaries to raise capital, and magically have no NPAs to speak of.

Cryptogeeks justifying 'underlying' whatever and using leverage from lenders who like 8-10%. Apparently there is an entire market where lending is happening to crypto stuff that is disguised as treasuries or other currencies. And this is lending is outside of the banking channels almost like the private credit mania. 

Companies getting equity capital and pre-IPO capital to keep bankrolling their acquisitions or ventures, while keeping negative ROE (something a lender would definitely not like)

The everything bubble is scary.

My wife asked me when I doubted a particular crypto: "But if its going up, and has gone up for 5 years or even 10 years, doesnt it mean that you are wrong or plain envious?"

I said: "Yes, it means that I have been wrong until now."

Akshay Khanna from Taal said it well : "Aaawaaz unchi karne se, ya chillane se, aapka jooth sachh nahi ban jaata"

But beyond whether i am right or wrong (I am right of course), the big factor is the interconnectedness of things.

The wealth effect, and the leverage effect is probably the most I have seen even if I count the IT bubble of 1990s.

I asked chatgpt this:

Here’s the “Magnificent 7” + Oracle + Broadcom market caps at 2019-12-31 vs 2025-09-17 (USD, billions). Sources are shown per row.

Company	2019-12-31	2025-09-17 (yesterday)
Apple	1,304.76	3,550.00. (StockAnalysis)
Microsoft	1,203.06	3,791.06. (StockAnalysis)
Alphabet (Google)	923.77	3,017.82. (StockAnalysis)
Amazon	916.15	2,470.21. (StockAnalysis)
NVIDIA	144.02	4,138.05. (StockAnalysis)
Meta Platforms	585.32	1,948.71. (StockAnalysis)
Tesla	75.40	1,420.00. (StockAnalysis)
Oracle	169.94	856.52. (StockAnalysis)
Broadcom	125.71	1,634.74. (StockAnalysis)

Group totals (Mag7 + Oracle + Broadcom)

• 2019-12-31: 5,448.13

• 2025-09-17: 22,827.11

• Absolute change: +17,378.98

• Multiple vs 2019: 4.19× (≈ +318.99%)

(Arithmetic shown using exact figures above.)

Notes on sources & dates

• All figures are end-of-day market caps from StockAnalysis pages that list both historical (Dec-31-2019) and current values for Sep-17-2025; Apple and Amazon pages explicitly show “as of Sep 17, 2025.” Tesla’s page shows Sep-17-2025 in the header (“Market Cap … as of September 17, 2025”) and the 2019 figure in the history table. (StockAnalysis)

So; basically, even from pre-covid highs, we have $17 T of wealth sloshing around the world economy just from these 9 companies.

I am quite scared of a 1987 crash kind of scenario...

A fat finger, or a Meta saying that capex will drop from the guided 70 B to only 35 B would lead to quite the avalanche. 

Again - I highlight the interconnectedness risk, and the euphoric wealth effect where social media is full of people who got the right stock, and right portfolio, etc.

"We may not know where we are going, but we sure as hell ought to know where we are" - Howard Marks

Where we are is close to the summit. 

And just like the 1989 Japan, or the 1999 IT bubble, the narrative is: "dude, the world is going to change around this new epicenter, and all we have to do is sit tight."

 

Fragile Markets

There is something really fragile about the top companies' makeup.

Oracle just moved to almost 1 T, Broadcom is at 1.7 T...

There are 4 big hyperscalers now; they all buy most of their stuff from 1 supplier - NVIDIA.

NVIDIA makes all of its stuff from 1 supplier - TSMC

TSMC buys most of its stuff from 1 supplier such as ASML, and the Japanese companies...

The hyperscalers are deeply entwined with particular companies.

OCI depends on OpenAI and Meta, and NVIDIA (as customers)

Azure depends on OpenAI

Google depends on itself and Anthropic

AWS has many customers?

All of them need nutty amounts of power.

If claude, or chatgpt begin losing customers (switching costs are super low now) - what happens?

__________

Then there is the pvt credit market which is also depending on steady inflows because the round robins seem to have started. Banks like JP M are simply stepping back...

___

And then, gold, bitcoin and crypto treasury companies.

____

Looks like a weird pack of cards

Something's off

Nov 2024: I wrote: "Boom, boom boom, and I fired those people, emptied the whole building, but not a whine... its like those people had never existed."

Somethings gotta give

And here we are in August 2025

• The China real estate bubble burst - and nothing happened? Nobody's upset, people havent lost their life savings? Normally, when RE is a big chunk of the economy, and if RE bubbles burst, there is degrowth - how has China continued to grow?
  This makes me believe that similar to how nobody expected degrwoth and stagnation in Japan in 1985; this time too, there is degrowth and its like the 4th dimension that nobody can see.
• Israel attacked Iran, and the nuke facilities got bombed and no retaliation?
• Nobody cares about Gaza
• People are bugged with Putin, but he just isnt backing down
• Trump announces tariffs on everybody and the US economy still grows?
• The Mag 7 turn from FCF to low FCF and yet they add trillions in market cap?
• The Pvt Credit bubble continues unabated, and nobody cares about NPAs?
• Crypto assets are now mainstream? I mean, it says "Crypto".

Its all very fishy.

Reminds me of when I read this book in 2015; and it took until 2022 or so for the bubble to burst.

Bad!

Include me out

 Something just feels off.

The private equity and private credit market are just too big man. 

And then, that the NVIDIA GPUs are a fast depreciating asset of not more than 2-3 years of real life + I think its a bit like quantum physics (nobody really gets it), which is why Deepseek did what??

Scenarios:

NVIDIA falls to $1.5 T

That small island is scared because the ships came too close? Wait, they make 90% of all chips?

All those AI factories are kind of obsolete, because AI figured out how to make its own reasoning and inferencing more efficient

50% Depreciation of the currency

Something weird about gold going up 40% in a year ?

Hmmm; 4% NPAs on private credit. Of course, there is no hidden borrowings and cross collateralisations

Whats up with Elon + Trump? What if they have a fall out?

Things just feel odd. 

Old Man Buffett is ready.