Are we in a bubble?

[Spekulatius]

1 hour ago, Paarslaars said:

Yeah but this time there are earnings though...

Cisco and JDSU, COHR had earnings too. Glas fiber , optical filters etc could be sold for exorbitant prices for a while. The demand went down in 2001/2002 but more importantly the prices went down even more so, so the revenue and margins went to the dogs for about 24 years until now basically.

[KCLarkin]

1 hour ago, Spekulatius said:

Cisco and JDSU, COHR had earnings too. Glas fiber , optical filters etc could be sold for exorbitant prices for a while. The demand went down in 2001/2002 but more importantly the prices went down even more so, so the revenue and margins went to the dogs for about 24 years until now basically.

 

A bit interesting to lump in Cisco and JDSU. Cisco was mostly a simple overvaluation story.

 

JDSU was a bullwhip effect story.

 

I do wonder if the AI “bubble” results in a bullwhip effect. And I think AI investors should read the JDSU” case study” so they are aware of the potential risk.

[dealraker]

5 hours ago, Spekulatius said:

Roughly 40 out of 500 stocks of the SP500 have driven all the returns and trade at high multiples. The rest not so much and many of the remainder trade at decade low multiples. It’s the same kind of bubble we saw in 1999 , when the market got top heavy to the extend it skewed the SP500. In fact some of the same stocks that went to sky high valuations in 1999 are back in the game again.

Yep

 

[Sweet]

5 hours ago, Paarslaars said:

Yeah but this time there are earnings though...

Risk is that future earnings of companies don’t materialise.  Anthropic and ChatGPT count for huge portions of the cloud backlog at Google and MSFT respectively.  If for something happens to either ChatGPT or Anthropic I can see a sharp contraction in many AI stocks.

[Mephistopheles]

13 hours ago, Paarslaars said:

Yeah but this time there are earnings though...

True but last time there was far less capex, at least  amongst the largest companies

[Parsad]

1 hour ago, Mephistopheles said:

True but last time there was far less capex, at least  amongst the largest companies

 

No, there was as much capex last time, if not more relative to cash flow...Nortel, 360 Networks/LVLT, Cisco, Amazon, etc. 

 

The other aspect was that simple access to the internet did not create efficiencies that fell directly to operating margins.  AI is doing that...which is very different than the Industrial revolution, Digital revolution or Internet revolution.  Because of the additional value from general artificial intelligence, it is having a direct influence on operating margins and creating real earnings that weren't present in other economic revolutions. 

 

Not saying AI isn't in a bubble...just that the size and scope is markedly different than past bubbles in terms of ACTUAL business improvements and efficiencies.  Cheers! 

[frommi]

3 hours ago, Parsad said:

The other aspect was that simple access to the internet did not create efficiencies that fell directly to operating margins.  AI is doing that...which is very different than the Industrial revolution, Digital revolution or Internet revolution.  Because of the additional value from general artificial intelligence, it is having a direct influence on operating margins and creating real earnings that weren't present in other economic revolutions. 

In theory, but are there examples where that is really happening?

[Mephistopheles]

6 hours ago, Parsad said:

 

No, there was as much capex last time, if not more relative to cash flow...Nortel, 360 Networks/LVLT, Cisco, Amazon, etc. 

 

The other aspect was that simple access to the internet did not create efficiencies that fell directly to operating margins.  AI is doing that...which is very different than the Industrial revolution, Digital revolution or Internet revolution.  Because of the additional value from general artificial intelligence, it is having a direct influence on operating margins and creating real earnings that weren't present in other economic revolutions. 

 

Not saying AI isn't in a bubble...just that the size and scope is markedly different than past bubbles in terms of ACTUAL business improvements and efficiencies.  Cheers! 

Perhaps I stand corrected, but just referring to say the largest 5-10 cos like MSFT, INTC, ORCL, etc vs the largest ones today. The capex was same (relative to fcf)? Didn’t realize CSCO’s capex was that high but it does make sense. I should’ve looked it up. 

[ajlomb1011]

I think this will come down to,

 

1. How many data centers are actually being built and if they’re even operational.

 

2. How much cash Anthropic and OpenAI are burning to run their models. 

3. LLMs eventually don’t improve anymore and we’ve been throwing trillions in compute at something that has hit a wall. 
 

honestly if any of those 3 things end up being true then I see the whole semi/AI train slowing and bringing the whole market with it. 

[Paarslaars]

1 & 2 are valid points. 

Not concerned with 3, this is scaling exponentially...

[Spooky]

6 minutes ago, Paarslaars said:

Not concerned with 3, this is scaling exponentially...

 

Can it really scale exponentially though if there are real world / physical constraints like building data center capacity / chips etc.?

[Paarslaars]

Well that is why I agree with the first two... money & data center capacity are the bottlenecks to AI progress. It's not like we are reaching physical or technological limits. There is no wall to hit on that front.

[frommi]

29 minutes ago, Paarslaars said:

1 & 2 are valid points. 

Not concerned with 3, this is scaling exponentially...

No, LLMs do not scale exponentially. In fact, the math governing Large Language Models dictates the exact opposite: they follow a power law of diminishing returns. 

Presented to you by Gemini

[Castanza]

On 5/16/2026 at 1:14 PM, Blake Hampton said:

It's kind of funny how offended some people get on here simply by hearing someone else's opinion. The internet is a bad place for that.
 

The math looks incredibly bad to me. Market bubbles are also largely behavioral, so when you find large groups of people harping on you for not being fully invested, while at the same time refusing to back their arguments with any sort of logic or fundamentals, I don't think it's a good indicator.
 

I'm investing my money in a way that'll do well in the future I see. Maybe you disagree and see the future differently than I do, and that's fine. That's what makes a market. But there's only one future, and we're all going to have to live through it. What's said here means nothing. What matters is being right over the long-term and making sure your money is invested in a way where you can profit from it.

 

LOL the only one throwing insults is you "wisdom from Trumpers" (idek what that means) and the only one being offended is you. People on this forum have shown they have skin in the game and have demonstrated many times valuations are attractive in many market segments and that with a truly long-term horizon being in the market is better than sitting on a bunch of cash or t-bills earning basically nothing when sized against inflation. 

 

My problem and many others problem is you come on an "investment" forum, lambast people on here for being invested saying everyone on here is a moron because we don't read the FT, NYT etc. and that we are about to have the worst fiscal crisis in history and everything is going to zero. Yet you never disclose your position, show your returns, show how you're positioned in alternative assets etc. If you're going to talk the talk then walk the walk and let everyone know how you're positioned and how we should position. Otherwise you're just a lunatic standing on the street corner squawking. 

 

I think you've been on here three years? Only disclosed some real estate and a position in OXY often citing Vicki as a great leader (LMAO she gone) or selectively quoting Buffett while completely ignoring the fact that Berkshire is actively buying back their own shares. Are they wrong too?

 

 

[Castanza]

47 minutes ago, frommi said:

No, LLMs do not scale exponentially. In fact, the math governing Large Language Models dictates the exact opposite: they follow a power law of diminishing returns. 

Presented to you by Gemini

 

We've also yet to see this play out from a business operational standpoint. Plenty of companies that are utilizing AI are struggling to properly and efficiently use tokens.  I think good business operators will be more important than ever. AI has made exploring new avenues easier than ever. But historically too many avenues is a drag on company performance. It will be interesting to see how this plays out but companies that lack focus struggle. 

 

A good case study on this is from the 1990s company General Magic. Some products ahead of their time, too many irons int he fire, not enough focus on the core product, too many rabbit trail features etc. I find some parallel to AI with this story as there are more and more stories of businesses not knowing what to do with AI or how to effectively apply it for positive business efficiencies. 

 

https://en.wikipedia.org/wiki/General_Magic

 

 

 

[Castanza]

@Blake Hampton 

 

Nasdaq 100...Last 20 years returns are 15% yr...earnings have compounded at 15% for the index over the same timeframe....If you bought right at the peak in the early 2000's endured that 82% drawdown and held to today you still compounded at 8.8%/yr for 26 years...

 

EPS quarter over quarter is 45% and looking like close to 30% on the year. Forward PE was recently at ~30x with 9 of the top 10 companies having PE below 40. Dotcom bubble had the index north of 100 forward PE and 9 of the top 10 companies had PE also north of 100. Today 20% of the index has a net margin of 50 and 100% in 1999 there were zero companies in that bucket. Today 50% of companies have margins between 25 and 50%. in 99 43% of index had margins 10-25% margins...21% of the index had 0-10% margins (vs 9% of index today) and 10% was negative....

[Gregmal]

Must be MAGAmath 

[Lazarus]

On 5/16/2026 at 1:14 PM, Blake Hampton said:

 

The math looks incredibly bad to me...
 

I'm investing my money in a way that'll do well in the future I see.

 

Which math is that? 

 

Are you brave enough to share exactly what you've invested? Both now and in the last 5 years?

[frommi]

2 hours ago, Castanza said:

@Blake Hampton 

 

Nasdaq 100...Last 20 years returns are 15% yr...earnings have compounded at 15% for the index over the same timeframe....If you bought right at the peak in the early 2000's endured that 82% drawdown and held to today you still compounded at 8.8%/yr for 26 years...

 

EPS quarter over quarter is 45% and looking like close to 30% on the year. Forward PE was recently at ~30x with 9 of the top 10 companies having PE below 40. Dotcom bubble had the index north of 100 forward PE and 9 of the top 10 companies had PE also north of 100. Today 20% of the index has a net margin of 50 and 100% in 1999 there were zero companies in that bucket. Today 50% of companies have margins between 25 and 50%. in 99 43% of index had margins 10-25% margins...21% of the index had 0-10% margins (vs 9% of index today) and 10% was negative....

Yes but the origin of all these earnings are the Capex spend from the Mag7. If we do this for 10 years the world will be taped with data centers. Do you really believe that will happen?

[Castanza]

6 minutes ago, frommi said:

Yes but the origin of all these earnings are the Capex spend from the Mag7. If we do this for 10 years the world will be taped with data centers. Do you really believe that will happen?

 

I have no idea what will happen. I'm not rushing to buy the Mag7 (I do own GOOGL with $150 cost basis). I think history shows that the market goes through bubbles and cycles pretty consistently and even if you get your timing 100% wrong and buy at the peak, you will generally do just fine and outperform cash, bonds, and real estate. 

 

There is always a bubble, always a cycle, always new fad companies, and neglected sectors. Yet unless history changes (as Blake claims) the market for the most part has always been up and to the right over a 20+ year time horizon. 

 

 

[Blake Hampton]

I think there's only one way to convince you guys:

 

"What Blake is doing with investments is frankly unbelievable, nobody has ever seen anything like it. We’re talking five thousand percent, ten thousand percent, maybe even a million percent returns in one day. The fake news won't report on it, but they’re the greatest in the history of our country, maybe ever. Other investors? Total losers, they're pathetic. People are calling me, very smart people, top people, and they’re saying, 'Sir, Blake’s numbers are too big, they're breaking the calculators.' Can you believe that? Breaking the calculators, it's true."

[Castanza]

3 minutes ago, Blake Hampton said:

I think there's only one way to convince you guys:

 

"What Blake is doing with investments is frankly unbelievable, nobody has ever seen anything like it. We’re talking five thousand percent, ten thousand percent, maybe even a million percent returns in one day. The fake news won't report on it, but they’re the greatest in the history of our country, maybe ever. Other investors? Total losers, they're pathetic. People are calling me, very smart people, top people, and they’re saying, 'Sir, Blake’s numbers are too big, they're breaking the calculators.' Can you believe that? Breaking the calculators, it's true."

 

[brobro777]

I remember watching this live on CNBC during the bubble days - https://en.wikipedia.org/wiki/Emulex_hoax

 

haha

 

 

 

[Malmqky]

If the market is outperforming my portfolio, it’s a bubble.

 

Simple really, you guys are overcomplicating it

[Malmqky]

22 minutes ago, Castanza said:

 

The problem with OXY is that Blake is viewing it with an ultra long timeframe, and his thesis technically won’t be proven wrong until decades in the future. Until then, the underperformance will mean nothing to him. The empire building and poor capital allocation means nothing to him.

 

Personally seems too hard to me. 3-5 year thesis gives you time to pivot. The 30 year thesis will kill you if wrong, and still might due to opportunity cost even if right.

Edited May 18 by Malmqky