I want to get to Nvidia in a second, but
we just had a really fascinating
conversation, not unrelated, with
Winston Ma, formerly the CIC about the
fact that he thinks this is a sovereign
wealth fund, uh, the US taking stakes in
Intel and MP materials and so forth. Do
you agree?
>> Morning, Matt. Well, I think it's a bit
more of a of a utility play than it is a
sovereign wealth fund play. Um but the
the I think the underlying message is
true which is uh we must build uh
refinery capa capacity and capability
inside of our own country.
>> So and you are uh obviously you're a
venture capitalist. Um we've talked to
you a lot in connection u with Mark
Douglas and uh Mountain um but you
worked a lot on AI and in the technology
sector that's kind of your your
wheelhouse. Do you think we need the US
to have either a sovereign wealth fund
or its own venture capital fund or or uh
a utility play of some sort in order to
uh support the national security of our
infrastructure?
>> Absolutely, Matt. I think, you know,
when you think about the alternative of
what happens with the pressures
associated with Taiwan, the global
political pressures associated with
Taiwan, the alternative is much worse
than us um having a national uh refining
and fab uh capacity.
>> And uh I'm taking a look at a recent
post that you make made on LinkedIn
about this uh potential intel investment
by the US government. You said that this
will depend on how President Trump is as
an investor. You also make the point
that the company, it does have the
technology to become a TSMC substitute
or at least the backup to TSMC. You also
write about the potential that the
government might be able to motivate
Nvidia and AMD to commit meaningful
volume to Intel when it comes to fab and
packaging. With all this in mind, Joe,
do you follow the US government into
Intel? Is Intel a viable investment from
where you're sitting?
>> Uh, great question, Katie. Good morning.
Um I think that first of all I think
Intel is an interesting play for the
government to support. Um as you look
about as you look across the landscape
yet from recent earnings uh obviously
Nvidia crushed earnings in uh TSMC did
AMD Broadcom and Intel missed and Intel
announced layoffs uh coupled with their
miss. So I think the business is too big
to fail and needs the government support
and one of the things the government can
do to support it is uh motivate them to
move into the a more of a third party
refining uh business model allah TSMC
coupled with bringing capacity and uh
demand from AMD and uh Nvidia. So so yes
I I think that's the right play for the
government to take from a from a public
market investor standpoint. It's not my
forte, so I'll defer to the to the long
onlyies and the hedge funds out there.
Uh, but I think from a government
standpoint, it is the right play from a
national security point of view.
>> Fair enough. Fair enough. I am curious
what you think about the potential
timeline here. Again, going back to your
post and what you're saying, it sounds
like Nvidia, or not Nvidia, Intel
rather, has all the correct ingredients
here. Now, it of course has some
government backing as well, but this
isn't like turning a light switch, Joe.
How long do you think that it might take
for some of these factors to really come
together?
>> Well, Kate, some interesting things are
a foot. So, TSMC as part of their
earnings announced 15 new fabs uh
distributed globally and they intend to
begin packaging in their Arizona
facility at some point in the future,
but today all the packaging occurs in
Taiwan. So, that is a critical supply
chain element that's at a geopolitical
risk today. So from if you're the
government and you're Intel, this is a
decadel long journey to to convert the
Intel business model to a a third-party
refining capability as well as building
the supply chain, the domestic supply
chain around it. So this is not a 12 or
24month transition. This is a decade
long project. But I think again from a
national security standpoint, it is one
well worth taking.
>> Dude, while we have you, I want to get
to Nvidia. I think you said they crushed
it. Um I don't know if the market shares
that view. Uh what's your take on what
happened yesterday with the numbers and
what the outlook signifies?
Matt, I think the street is missing uh
the the concern over the data center
revenue. You know, if if you look at
what is happening from in their data
center revenue business, it is not a
demand problem. It is a supply problem.
They are the their black belt chips are
sold out well into next year. TSMC
capacity is sold out into next year. So
the the challenge is not how do they
generate demand and get orders. The the
challenge that they face is how do they
make chips fast enough to to fulfill
these orders. So the fact that Nvidia
was able to come in line with revenue
forecast despite the the bottlenecks in
supply chain I think is is really
compelling. You couple that with, for
instance, the the earnings report that
came out of Core Weeb and seeing them
continue to grow at 200% year-over-year
with Evid margins north of 60%. I'm I'm
had the luxury of investing in a company
called Lambda Labs and their performance
um that I I agree with Dan Ice like
we're in the very first innings here.
What what do you think Joe and I ask you
as someone who has the perspective of
starting businesses uh working in this
sector? You know how long uh and how
much investment it takes to get
something off the ground. But if China
doesn't want its companies buying any
more Nvidia chips, can they just steal
the IP and make something on the level
of Blackwell? Or is it far too
technically challenging to get there?
It's far too technically challenging in
the short run, Matt. And I think what's
happening is um the the Huawei chip in
some respects is comparable to the H20.
You know, it has much it has much more
powerful compute, but is a much bigger
form factor. So, there's some trades
there, but in a in a export ready
version of the Blackwell chip, it is
leaps and bounds ahead of uh the
domestic Chinese uh alternatives. And
then if you think about what is in the
future with the with the Reuben chipset
coupled with TSMC's migration to the uh
to the uh 2nm production cap capability
the the gap is extending. Hey.