AI bubble

Does anyone else think we’re at the peak of an AI bubble? I think the fundamental problem is that the main reason for the sky high valuations of openAI is based on the expectation that they are going to create AGI - so no matter how much value they deliver with their new ChatGPT features and b2b infrastructure they cannot live up to the expectations of investors. Also AGI will probably not be here for at least another few decades.

How could we profit off this?

32 Comments
 

IB_Applicant

Do you have any plans to profit off this artificial demand?

I’m genuinely not sophisticated enough. Know your limits. 

I do own NVIDIA from when it was like $3-$5 a share though…

Commercial Real Estate Developer
 

PEarbitrage

I love people who have no idea what the fuck they are talking about making outrageous predictions on a market. 

I bet you were really big into NFTs. 

How much money would you say you spent on “Bored Apes?” 

...but is it REPE?
 

Can someone more plugged in walk me through the Oracle news this week?
- signed 455bn of long term contracts (unclear to me how deployed this is / yield earning)
- generating 900 million in sales
- converting into $125 million of net income (2.7% ROA)

So is the payback period 37 years? Will the chips they’re renting out even be relevant in 5 years? Who bears that capex risk?

 

https://x.com/brewmarkets/status/197533144711341308
PTJ says it feels like Oct 1999.


If that’s true markets haven’t even peaked yet. I think more companies that go public or announce big acquisitions might do the trick.

Wrt to how the bubble formation & crashing… - probably pvt companies cascading into public ones for valuation. Honestly even now OpenAI is paying too much on acquisitions and investors are paying too much for mediocre companies.

Rn I don’t see too much over-valuation on public large caps but there are too many inflated egos around in private side and investors are eating it up. 

Also most AI companies are tech stack and infra companies not AI application  companies… which means there’s also a risk of some revenue that these guys have being frothy. There’s a lot of mediocre AI applications built using really powerful tools and it's only going to spread more. AI hustle culture is unreal rn and there's not much quality control cuz companies all wanna make revenue and raise more money. Not even reputable funds like Sequoia can escape this.

When in doubt, use more peanut butter
 

Is there a single dark GPU? A key driver of the late 90s bubble was the huge amount of dark fiber that was being installed. The amount was so large that the ensuing bankruptcies in the early 00s led to a litteral loss of these assets. The last time I personally heard about dark fiber being found was in 2017.  According to older people in the industry than myself they believe that there might still be 15 - 18% of the fiber install in the late 90s that still is not found and activated. 

The modern buildout has no dark infrastructure. This is a fundamental difference that people for the most part are not talking about. Are there spreads between income and costs? Yes, but that is a bridge that tech firms have been able to overcome for more than 50 years. I don't see this as any different. Is it possible there is a retraction? Yes, but people who think this is 1999 again are too hindered by their past experiences. Don't forget the tech bubble has been at 1999 for 10 years according to market predictors. 

 

There’s no bubble yet. But it’s frothy. 

AI revenue is slowly and slowly getting frothy and we’re couple IPOs and acquisitions that make no sense away from a bubble IMO. 

That’s what they mean by “it’s 1999”. By most metrics, there was no bubble in October of 1999. P/E ratios are ok right now but revenue is getting frothy. Madness hasn’t started.


OpenAI almost bought windsurf and if they did they’d have eaten major shit. Signs are already showing for what’s very likely to come.

When in doubt, use more peanut butter
 

What made the late 90s early 00s a bubble was the complete divorce of infrastructure development and companies that had zero construtive way of delivering their proclaimed purpose with the actual consumption demand.  As I have said before gray hairs of the telecom industry believe that there is still as much as 18% of the dark fibre that was built during the boom that is still lost to this day. The market then was building as if 2001 was really 2021 in terms of internet capabilties and demand. 

There is not a single dark GPU in the entire data center market. Okay I should clarify this, there are dark GPUs that are there for redundancy, but there is zero excess compute capacity in the system now.  Is it possible that the hype will create a situation like the late 90s? It is possible, but I believe that the consumption demand is driving this development wave. We have major tech firms that are purchasing compute demand for 10+ years into the future with cashflow.  That is very different than a speculative IPO bubble

 

I think the speed of AI and also energy consumption of AI will be big places to improve.

"If you always put limits on everything you do, physical or anything else, it will spread into your work and into your life. There are no limits. There are only plateaus, and you must not stay there, you must go beyond them." - Bruce Lee
 

I am completely opposite of you on this. Efficency can be made better, but really AI could get slower if anything. Better outcomes are far more important than energy and speed of delivery. 

 

you're over generalizing. Ofc there are AI companies that are pure hype but don't add value to companies and consumers but there are AI technologies that can improve efficiency and productivity significantly in certain industries.

 

Fear the Beard

you're over generalizing. Ofc there are AI companies that are pure hype but don't add value to companies and consumers but there are AI technologies that can improve efficiency and productivity significantly in certain industries.

Perfectly vague response that uses a lot of words to say nothing and hedges everything. 

Who needs LLMs? We already have humans doing the same thing. 

Commercial Real Estate Developer
 

Velit et eveniet quis quia. Omnis velit ab harum in. Nemo consectetur excepturi quas natus occaecati beatae nesciunt.

Career Advancement Opportunities

June 2026 Investment Banking

  • Evercore 01 99.4%
  • Moelis & Company 01 98.8%
  • JPMorgan 01 98.2%
  • Guggenheim Partners 01 97.7%
  • Morgan Stanley 07 97.1%

Overall Employee Satisfaction

June 2026 Investment Banking

  • Moelis & Company No 99.4%
  • Morgan Stanley 01 98.8%
  • Evercore 01 98.2%
  • BMO Capital Markets 12 97.6%
  • Banco Santander 01 97.1%

Professional Growth Opportunities

June 2026 Investment Banking

  • Moelis & Company No 99.4%
  • Evercore No 98.8%
  • Morgan Stanley 05 98.2%
  • JPMorgan No 97.7%
  • BMO Capital Markets 12 97.1%

Total Avg Compensation

June 2026 Investment Banking

  • Vice President (14) $434
  • Associates (43) $259
  • 3rd+ Year Analyst (8) $210
  • 2nd Year Analyst (22) $179
  • Intern/Summer Associate (13) $156
  • 1st Year Analyst (75) $151
  • Intern/Summer Analyst (66) $101
notes
16 IB Interviews Notes

“... there’s no excuse to not take advantage of the resources out there available to you. Best value for your $ are the...”

Leaderboard

1
redever's picture
redever
99.2
2
Secyh62's picture
Secyh62
99.0
3
BankonBanking's picture
BankonBanking
99.0
4
kanon's picture
kanon
99.0
5
DrApeman's picture
DrApeman
98.9
6
dosk17's picture
dosk17
98.9
7
CompBanker's picture
CompBanker
98.9
8
GameTheory's picture
GameTheory
98.9
9
Betsy Massar's picture
Betsy Massar
98.9
10
numi's picture
numi
98.8
success
From 10 rejections to 1 dream investment banking internship

“... I believe it was the single biggest reason why I ended up with an offer...”