always good to have you in the Lion
City.
>> So when you take a look at the AI play,
given that China's AI players are much
cheaper, valuations are way cheaper than
the US, why wouldn't you be buying into
China, then try to keep on, you know,
wrecking up gains in uh in the US?
>> I think I think the rally certainly
started in the US, but it has brought it
out to areas like China because the
applications around AI are really
starting to expand. But let's not forget
what's happening in the US markets as
well. The US markets historically grew
earnings at about six or 7% a year and
post AI since 2022 US markets earnings
EPS is growing at double digits. So
we've grown from market that's growing
at 6 or 7% to now a market that's
growing at 11 to 12%. And for the next
three years we think that earnings
growth is once again going to be in the
double digits. And so you've seen a
little bit of an expansion in the
absolute valuation in the US but we
think it's much more than justified.
>> The thing is Nvidia disappointed. Was it
a surprise to you? I mean expectations
were skyhigh. It was kind of bound to
disappoint.
>> Yeah, look the stocks doubled in a very
short period of time. So expectations
had certainly gotten a lot higher but I
think it was a very very good good
result and fundamentally the question
around Nvidia is always what happens to
the trajectory of capex from these
levels right capex remember at the
beginning of the year capex the four
major hyperscalers were expected to be
around $280 billion you know what that
number is now it's $360 billion and so
that's the single biggest KPI that's
still impacting Nvidia and we think that
models are going to get increasingly
more competitive. Kyperscalers are
chasing after artificial general
intelligence. They're not happy with
standing still right now. They
understand that this is going to be the
biggest technology pivot that we've seen
in our lifetime and they want to be the
leaders and that race is not slowing
down and until it does, you know, Nvidia
is going to continue to be interesting.
>> The thing is, show me the money. If
capex is so high, I want to see the
returns.
>> I'm not seeing the kind of returns that
I should be seeing if I'm putting that
much money.
>> Yeah. So, you might start. So part of
the reason why Alibaba has been so
strong today is because they're actually
starting to see the returns around AI as
well. You're seeing a lot of consumer
internet platforms
>> at the level that you're investing
though.
>> So one one uh in interesting stat that
um Nvidia put out over the last couple
of quarters is that if you look at
training versus inference workloads
inference is actually the monetization
piece. Training is actually the building
of the models, right? We're now crossed
over to the point where more than 50% of
the workload is on inference relative to
training. And so you're actually getting
to the point where modernization is
increasing. And we're optimistic as we
look into 2026 with things like the
development of AI agents, the continued
monetization of consumer internet
platforms like Meta, Netflix, and
Spotify. The reason why these stocks
have been so strong is because they're
using AI and leveraging AI to actually
drive higher levels of monetization. So
there might not necessarily be visible
forms of ROI to us right now because
enterprise is just getting started, but
it's it's well underway. But you can
understand why people are beginning to
think about it being a bubble, right?
Because of the surge in the stock
prices. I mean,
>> you're optimistic about the future, but
you have changed your mind before. That
deep sick moment did prompt you to
reassess perhaps your position uh on
Nvidia. What would cause you to rethink
again?
>> Yeah, look, I I think what what would
cause us to rethink and first of all,
we've been talking about an AI tech
bubble for the last 10 years, right?
2015, it's oh, it's a bubble. 2016 it's
a bubble bubble though.
>> And by the way, the reason why is
because we had a such a sharp correction
down that this rally up has been so
powerful that people are saying, "Hey,
we need to be cautious." But if you
think about 2025 in isolation, the
NASDAQ is up 12%. Over the last 10
years, the NASDAQ has averaged about a
16% return per year. So, we're just
really just back on track to where it
has been, right? Um and if you told me
at the start of the year that the single
most in uh important KPI for AI trade is
AI capbacks and AI capbacks has gone
from 280 billion to 360 billion and now
the trajectory going forward is many
many people think that it's going to be
more than the size of the US Department
of Defense spending which is crazy to
think about. Then you you got to think
that the trade continues to have legs.
>> So how do you play this? I mean you buy
the dip. Uh yeah, I mean this is this is
a period of digestion. So like stocks
anytime
>> every dip, buy every dip,
>> you know, you buy every dip. But so you
have to think about what's happening and
you have to really go into the mindset
of these people that are the companies
that are spending this large amounts of
capital. So if you think about what Meta
is thinking, Meta is just hiring
engineers for hund00 million per
engineer. Is that because they're going
to stop spending on AI tomorrow? No. You
have to think about what they're chasing
after. And the big goal of what they're
trying to chase after is AGI. And AGI is
trying to unearth something beyond what
humans are capable of today. So, how do
we cure cancer? How do we solve climate
change? How do we solve like are there
aliens in existence, right? There's like
things that human intelligence can't
currently comprehend. And we're trying
to move towards that AGI and we're just
cracking the surface on that. And I just
don't see them slowing down in terms of
that. So outside of any kind of big
multi-standard deviation event where we
lead to a big correction. I just think
that the medium to long-term trajectory
of spend is upward, not downward. And so
for those reasons, I'd probably be
buying a lot of the dips.
>> I'm a little bit nervous because I think
there are risks there and I guess risk
are being underpriced right now. How do
you hedge? I mean, you know, just in
case you're wrong.
>> Yeah. Yeah. So look, I I think you have
to think about it in and one of the
things about Goldman is what we tried to
pride ourselves on is trying to be
humble and anytime the fact pattern
changes, we're going to change our
opinion as well, right? Uh at least the
investing culture is is that way. And
so, you know, we're going to continue to
look for signp points that maybe we are
wrong on this,
>> including like what what signposting? Uh
so basically anytime that we think
there's indicators that the capex
trajectory is slowing, think about
that's the most important KPI right now
in in the whole AI capex trade. And so
what we're going to look for is any
indication that that's slowing