AI Hype Triggers Tech Stock Decline
Key Points
- Tech giants like Nvidia, Microsoft, and Meta saw sharp pre‑market declines as investors grew nervous about AI‑related risks and competition.
- Apple’s App Store surged with the DeepSeek app, a free ChatGPT‑style chatbot that vaulted to the top ranking and sparked trader panic.
- analysts question how DeepSeek can handle massive user demand with the modest hardware it claims to use, suggesting hidden or sophisticated compute strategies.
- Many investors mistakenly believe that a copyable model would eliminate the need for costly chips, but true competitive advantage lies in continuous innovation, not model secrecy.
- The underlying belief is that demand for AI intelligence is essentially limitless, ensuring sustained market opportunities for companies that can deliver advanced capabilities.
Full Transcript
# AI Hype Triggers Tech Stock Decline **Source:** [https://www.youtube.com/watch?v=eHn2dIJWRbk](https://www.youtube.com/watch?v=eHn2dIJWRbk) **Duration:** 00:05:21 ## Summary - Tech giants like Nvidia, Microsoft, and Meta saw sharp pre‑market declines as investors grew nervous about AI‑related risks and competition. - Apple’s App Store surged with the DeepSeek app, a free ChatGPT‑style chatbot that vaulted to the top ranking and sparked trader panic. - analysts question how DeepSeek can handle massive user demand with the modest hardware it claims to use, suggesting hidden or sophisticated compute strategies. - Many investors mistakenly believe that a copyable model would eliminate the need for costly chips, but true competitive advantage lies in continuous innovation, not model secrecy. - The underlying belief is that demand for AI intelligence is essentially limitless, ensuring sustained market opportunities for companies that can deliver advanced capabilities. ## Sections - [00:00:00](https://www.youtube.com/watch?v=eHn2dIJWRbk&t=0s) **AI Panic Triggers Tech Stock Dip** - The speaker links recent AI developments—particularly DeepSeek’s free ChatGPT‑like app on Apple’s App Store—to today’s sharp declines in Nvidia, Microsoft, Meta, and other tech stocks. ## Full Transcript
we're having a grown-up moment for AI
today it's January 27th and we are
making the stocks go down people
sometimes ask me to give stock advice
and I don't and this isn't stock advice
but Nvidia is down 11% pre-market which
is kind of a big deal considering
they're in the trillions for
valuation and Microsoft is down like 4%
meta's down everybody is losing today in
the stock market why why are tech
companies taking a beating it's because
of AI and specifically because of a
company that can't launch an AI feature
that's right a company I have not
mentioned which seems to be doing better
than most ironically it's
Apple Apple's app store is still the
place where ordinary stock Traders go to
find out what is going on with AI I
guess because deep seek hit number one
in the Apple App Store over the weekend
and that panicked a lot of Traders
because effectively what deep seek
offers is chat GPT for free it offers
chat GPT for free it's quite good the
model has personality they claim it was
basically trained on a couple of
Chromebooks and I've read the white
paper and by the way I don't say that to
denigrate because it is incredible
mathematics and incredible engineering
work so really well done the question
remains whether or not you believe that
they trained it on such a light set of
chips how on Earth are they keeping up
with the spike in demand for compute
that goes with millions and millions of
people supposedly downloading your app
to try it no one has been talking about
issues with deep seek serving all of
those new requests they are getting some
chips from somewhere the story is
absolutely more complex than you think
it is perfectly possible for this to be
an incredibly well-engineered model that
advances state-of-the-art in a lot of
ways and also it may be hiding chips it
can be both so at the end of the day I
think what ordinary investors are
worried about who don't sit here and
stare at AI all day is they think that
if this thing used fewer chips we are
going to need less chips in the future
they also think that if a model can be
copied it means that the boat that they
are betting that these large companies
are building with their technical
Investments isn't OTE and look I got to
tell you if you thought that you could
have a moat with a model you don't know
AI the moat is the Innovation Edge the
moat is not a model that no one can ever
copy ever it's not how it
works but apparently it's how Wall
Street thought it worked
no the reason why you invest in AI or
not and this is not investment advice
this is just sort of an
observation you invest in AI if you
believe the rocks have begun to think
and there is essentially an infinite
demand for
intelligence if there's an infinite
demand for intelligence it does not
matter whether more intelligence becomes
available you will always want more of
it and somewhere somehow there will be
plenty of market for companies to have
pricing power with
intelligence if you don't think there's
an infinite demand for intelligence then
you are going to assume that we will
have no demand for further intelligence
at some point and a company like deep
seek will just soak up the market with a
free offering deep seek has to make
money too at some point
like I I guess to me I look at this I'm
like the only way deep seek as a model
existed is because other models
delivered state-of-the-art Chain of
Thought reasoning built on immense
compute and I assure you given the cash
bur open AI is facing if there was a
magical way to do this with less chips
they would have done
it and now that deep seek has op sourced
their entire approach that's great for
everybody because it means that
everybody can learn from all of these
advances uh Mark Zuckerberg reportedly
went into founder mode over the weekend
because he was so mad about sort of meta
and llama and falling behind on deep
seek so don't worry they will take all
the learnings they can from that paper
and then they will put their billions of
dollars to
work um but netn net
at the end of the day billions of
dollars are on the table because
advancing the state-ofthe-art one
requires a lot of chips and two then
requires a lot of chips to
serve if you make a request for
inference
compute it you're you're buying server
time Sam Alman has said they lose money
on 01 pro at $200 a month because people
use it so much that's how much it costs
I think that Jensen will be fine I
really
do um so we'll see this none of this is
investment advice but it's sort of a
moment when AI is big enough to make
multi-trillion dollar companies take a
hit and that's what we're seeing today
all right who knows where it's going
next