Everyone suddenly decided AI is "the new dot-com bubble" because Burry bought puts on NVDA and some other AI names. I get the concern. Some of these stocks have gone vertical. A nasty correction can absolutely happen.

But calling the whole AI thing a bubble – as in "this is fake, like Pets.com" – feels way off to me.

This is how I see it.

1. Burry is smart, but he isn't the final boss of markets

The guy nailed the housing crash. Respect forever for that.

But his default mode is basically:

• things are overvalued
• crash is coming
• just wait

If you are always early and always bearish, you will eventually be right on something. In the meantime, you can be wrong – and very early – for years.

Also, he is short specific stocks, not physics. Shorting NVDA or PLTR is not the same as saying "AI itself is nonsense and will vanish."

2. This is not 1999 internet

Back in the dot-com days, a lot of companies had:

• no product
• no users
• no real revenue

Now with AI we already have:

• companies doing multiple billions in annualized revenue almost entirely from AI products
• new AI tools going from 0 to hundreds of millions in ARR in 1 or 2 years
• actual customers paying real money because the stuff saves them time or makes them money

That’s a huge difference. This is not "maybe someday we’ll monetize attention". The monetization is already happening.

3. The “AI machines depreciate” argument

One of the big bear points is:

“All these AI data centers are going to depreciate in a few years. Everyone is overspending on GPUs that will be useless soon.”

Couple of simple points:

• Every kind of compute depreciates
• Cloud servers, storage, networking – all of it has a 3-5 year life on the books
• Nobody called AWS a “bubble” because servers depreciate

AI chips are just powerful compute. They don’t run one hype product and then die. They get used for:

• training new models
• fine-tuning company specific models
• running all the day to day inference: search, recommendations, copilots, fraud checks, etc.

When new hardware shows up, older chips don’t magically go to zero. They slide down the stack into cheaper, less latency-sensitive workloads.

Could there be some overbuild? Sure. That happens in every big tech cycle. But overbuilt infra (fiber, cloud, railroads, whatever) usually becomes the foundation for the next decade of growth.

4. Adoption is way deeper than most people realise

This is the part that makes me think AI is actually underhyped if you zoom out.

AI is already quietly doing stuff like:

• drafting emails and documents
• helping with code and debugging
• summarising long PDFs, calls, and meetings
• handling a chunk of customer support
• translating content across languages

ChatGPT going viral wasn’t just a meme. It went from 0 to tens of millions of users extremely fast because normal non-tech people actually got value out of it.

And the newer models are getting much better in "soft" ways that don’t show up in benchmarks but matter in real life.

Example from my own house:

• earlier versions of ChatGPT struggled to understand my 4 year old daughter’s kid-speech
• now it basically understands everything she says and answers her like a patient adult

That is not “hype narrative”. That is what it feels like when capability compounds.

5. Quiet changes: translation, video, boring enterprise stuff

Most of the AI talk is about image generators, chatbots, and NVDA.

Some huge changes are happening almost in the background:

Translation and video

Big platforms are already using AI to:

• auto dub videos into multiple languages
• auto translate captions and comments

That means:

• a video recorded in one language can suddenly reach audiences in many other languages
• the amount of watchable content in your own language explodes

That is a long term boost to engagement, ad inventory and revenue that is hard to see day to day but real.

Enterprise workflows

AI is getting baked into:

• Office / Google Docs
• CRMs and ERPs
• code editors
• support tools, ops tools, internal search

This is the boring stuff that actually prints money over decades. Once companies wire this in and it works, they don’t rip it out. They build more on top of it.

6. Yes, some valuations are crazy. That’s different from “AI is fake.”

Two separate questions:

1. Are some AI linked stocks overvalued?
   • very likely yes
2. Does that mean AI itself is just a bubble like tulips or Beanie Babies?
   • I don’t think so

Dot-com is the obvious analogy:

• the internet was real
• a lot of internet stocks were not

If you bought pets dot com at the top, you got destroyed. If you held Amazon or Google for 10+ years, you did fine.

AI is probably going to look similar. Plenty of junk, a few monsters that become part of the basic infrastructure of everything.

7. My simple take

I’m not saying "all in AI at any price". I’m not saying “Burry is dumb”.

I’m saying this:

• AI is already woven into daily life in a way most people underestimate
• the money is already real, not theoretical
• companies and governments are just getting started on the heavy spending
• there will absolutely be corrections, maybe big ones, but that doesn’t make the whole thing fake

So I ask myself two dumb questions:

• in 10 years, will AI be more deeply integrated into how we work, create and communicate than it is today?
• in 10 years, will the total revenue and profits around AI be bigger than today?

If the answer is yes to both, then I don’t see “AI is a bubble” as a reason to avoid the entire thing. It just means you have to be ready for volatility.

Obviously not advice, just how I’m thinking about it while everyone keeps screaming “bubble” every time there is a red day.